!git clone https://github.com/syedmehran/cmps3160-project.git
%cd cmps3160-project/

Cloning into 'cmps3160-project'...
remote: Enumerating objects: 570, done.
remote: Counting objects: 100% (165/165), done.
remote: Compressing objects: 100% (93/93), done.
remote: Total 570 (delta 101), reused 131 (delta 72), pack-reused 405
Receiving objects: 100% (570/570), 688.84 MiB | 18.42 MiB/s, done.
Resolving deltas: 100% (331/331), done.
Updating files: 100% (353/353), done.
/content/cmps3160-project

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import glob
import yfinance as yf
import datetime

#Load stock data in

financ_UAL = yf.download('UAL', start = '2015-04-30') # united
financ_ALK = yf.download('ALK', start = '2015-04-30') # alaskan
financ_AAL = yf.download("AAL", start = '2015-04-30') # american
financ_ALGT = yf.download("ALGT", start = '2015-04-30') # allegiant
financ_HA = yf.download("HA", start = '2015-04-30') # hawaiian
financ_LUV = yf.download("LUV", start = '2015-04-30') # southwest
financ_SAVE = yf.download("SAVE", start = '2015-04-30') # spirit
financ_JBLU = yf.download("JBLU", start = '2015-04-30') # jetblue
financ_DAL = yf.download("DAL", start = '2015-04-30') # delta
financ_JETS = yf.download("JETS", start = '2015-04-30') # JETS etf index
financ_FRONTIER = yf.download("ULCC", start='2015-04-30')
financ_OIL = yf.download('CL=F', start = '2015-04-30')

financ_airlines = {
    'United': financ_UAL,
    'Alaskan': financ_ALK,
    'American': financ_AAL,
    'Allegiant': financ_ALGT,
    'JetBlue': financ_JBLU,
    'Hawaiian': financ_HA,
    'Southwest': financ_LUV,
    'Spirit': financ_SAVE,
    'Delta': financ_DAL,
    'Frontier': financ_FRONTIER,
    'JETS' : financ_JETS
}

airline_colors = {
    'United': '#1f77b4',  # muted blue
    'Alaskan': '#2ca02c',  # cooked asparagus green
    'American': '#d62728',  # brick red
    'Allegiant': '#ff7f0e',  # safety orange
    'JetBlue': '#9467bd',  # muted purple
    'Hawaiian': '#e377c2',  # raspberry yogurt pink
    'Southwest': '#7f7f7f',  # middle gray
    'Spirit': '#bcbd22',  # curry yellow-green
    'Delta': '#17becf',  # blue-teal
    'JETS': '#8c564b',  # chestnut brown
    'Frontier': '#c7c7c7'  # silver gray
}

[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed
[*********************100%%**********************]  1 of 1 completed

# Load in airline financial data
financials_2023 = pd.read_csv('../cmps3160-project/finance-data/T_F41SCHEDULE_P12_2023.csv')
financials_2022 = pd.read_csv('../cmps3160-project/finance-data/T_F41SCHEDULE_P12_2022.csv')
financials_2021 = pd.read_csv('../cmps3160-project/finance-data/T_F41SCHEDULE_P12_2021.csv')
financials_2020 = pd.read_csv('../cmps3160-project/finance-data/T_F41SCHEDULE_P12_2020.csv')

financials_2020.tail()

financials = pd.concat([financials_2023, financials_2022, financials_2021, financials_2020], ignore_index=True)

for airline, df in financ_airlines.items():
    print(f"{airline} DataFrame size: {df.shape}")
    print(f"Null values in {airline}:")
    null_counts = df.isnull().sum()  # counts of null values in each column
    print(null_counts[null_counts > 0])  # print columns with null values
    print("\n")

print(financ_OIL.isnull().values.any())
financ_OIL.shape

United DataFrame size: (2173, 6)
Null values in United:
Series([], dtype: int64)


Alaskan DataFrame size: (2173, 6)
Null values in Alaskan:
Series([], dtype: int64)


American DataFrame size: (2173, 6)
Null values in American:
Series([], dtype: int64)


Allegiant DataFrame size: (2173, 6)
Null values in Allegiant:
Series([], dtype: int64)


JetBlue DataFrame size: (2173, 6)
Null values in JetBlue:
Series([], dtype: int64)


Hawaiian DataFrame size: (2173, 6)
Null values in Hawaiian:
Series([], dtype: int64)


Southwest DataFrame size: (2173, 6)
Null values in Southwest:
Series([], dtype: int64)


Spirit DataFrame size: (2173, 6)
Null values in Spirit:
Series([], dtype: int64)


Delta DataFrame size: (2173, 6)
Null values in Delta:
Series([], dtype: int64)


Frontier DataFrame size: (682, 6)
Null values in Frontier:
Series([], dtype: int64)


JETS DataFrame size: (2173, 6)
Null values in JETS:
Series([], dtype: int64)


False

(2172, 6)

financ_airline_codes = {'United': 'UAL', 'Alaskan': 'ALK', 'American': 'AAL',
                        'Allegiant': 'ALGT', 'JetBlue': 'JBLU', 'Hawaiian': 'HA',
                        'Southwest': 'LUV', 'Spirit': 'SAVE', 'Delta': 'DAL',
                        'JETS': 'JETS'}

flight_airline_codes = {'American':'AA', 'Delta':'DL', 'Southwest':'WN',
                        'United':'UA', 'Alaskan':'AS', 'Spirit': 'NK',
                        'Hawaiian':'HA', 'Allegiant':'G4', 'JetBlue':'B6',
                        'Frontier':'F9'}

financ_JETS.head() # displaying JETS ETF to check data and columns

plt.figure(figsize=(14,7))

for airline, df in financ_airlines.items():
    plt.plot(df['Close'], label=airline, color=airline_colors[airline])

plt.title('Time Series Graph of Closing Prices ($)')
plt.xlabel('Date')
plt.ylabel('Closing Price')
plt.legend()

plt.show()

for airline, df in financ_airlines.items():
    df['Daily Return'] = df['Close'].pct_change().dropna()
financ_OIL['Daily Return'] = financ_OIL['Close'].pct_change().dropna()

combined_returns = pd.DataFrame()

for airline, df in financ_airlines.items():
    combined_returns[airline] = df['Daily Return']

mean_returns = combined_returns.mean()
cov_returns = combined_returns.cov()
std_returns = combined_returns.std()
combined_returns.dropna(inplace=True)
combined_returns.head()

correlation_matrix = combined_returns.corr()

import seaborn as sns

#create correlation matrix
cmap = sns.diverging_palette(10, 130, as_cmap=True)
plt.figure(figsize=(10, 8))
sns.heatmap(correlation_matrix, annot=True, cmap = 'Blues', vmin=-1, vmax=1,
            xticklabels=correlation_matrix.columns,
            yticklabels=correlation_matrix.columns)
plt.title('Correlation Matrix of Daily Returns')
plt.show()

# combine oil returns with airline returns
financ_OIL['Daily Return'] = financ_OIL['Close'].pct_change()
combined_corr_data = financ_OIL[['Daily Return']].rename(columns={'Daily Return': 'Oil Return'})

for airline, df in financ_airlines.items():
    combined_corr_data[airline] = df['Close'].pct_change()
combined_corr_data.dropna(inplace = True)

combined_corr_data.head()

delays = [7, 14, 30,60,90,180,365]
airline_names = ['United', 'Alaskan', 'American', 'Allegiant', 'JetBlue', 'Hawaiian', 'Southwest', 'Spirit', 'Delta', 'JETS']

correlation_results = pd.DataFrame(index=airline_names, columns=delays)

for delay in delays:
    for airline in airline_names:
        shifted_returns = combined_corr_data[airline].shift(-delay)
        correlation = combined_corr_data['Oil Return'].corr(shifted_returns)
        correlation_results.at[airline, delay] = correlation

correlation_results.head(10)

import scipy.stats as stats


significance_results = {}
n = 2160

correlations = {
    'United': [-0.049228, -0.030873, -0.025656, -0.079161, -0.012919, -0.016999, -0.004876],
    'Alaskan': [-0.041154, -0.019935, -0.021779, -0.066841, -0.004358, 0.001154, -0.000124],
    'American': [-0.056931, -0.016334, 0.00937, -0.070561, -0.003926, -0.005263, 0.007494],
    'Allegiant': [-0.030043, -0.073951, -0.046508, -0.029428, -0.025044, -0.018937, 0.008539],
    'JetBlue': [-0.061789, -0.036663, -0.01862, -0.05643, -0.004846, 0.003824, 0.003574],
    'Hawaiian': [-0.058751, -0.029893, -0.040023, -0.045276, 0.003318, -0.037189, 0.001015],
    'Southwest': [-0.042116, -0.020714, -0.035552, -0.045155, 0.006958, -0.007177, 0.000559],
    'Spirit': [-0.029643, 0.01245, -0.048053, -0.058677, -0.006752, 0.001502, 0.004846],
    'Delta': [-0.068438, -0.004271, -0.015327, -0.060281, -0.007192, -0.008411, -0.003624],
    'JETS': [-0.061341, -0.02731, -0.032695, -0.061719, -0.007318, -0.005128, -0.003519]
}


for airline, corr_values in correlations.items():
    significance_results[airline] = {}
    for delay_index, r in enumerate(corr_values):
        df = n - 2
        t_stat = r * (df ** 0.5) / ((1 - r ** 2) ** 0.5)
        p_value = 2 * (1 - stats.t.cdf(abs(t_stat), df))
        significance_results[airline][delays[delay_index]] = p_value

# Display the p-values
for airline in significance_results:
    print(f"Airline: {airline}")
    for delay in significance_results[airline]:
        print(f"Delay {delay} days: p-value = {significance_results[airline][delay]:.4f}")

Airline: United
Delay 7 days: p-value = 0.0221
Delay 14 days: p-value = 0.1515
Delay 30 days: p-value = 0.2333
Delay 60 days: p-value = 0.0002
Delay 90 days: p-value = 0.5484
Delay 180 days: p-value = 0.4297
Delay 365 days: p-value = 0.8208
Airline: Alaskan
Delay 7 days: p-value = 0.0558
Delay 14 days: p-value = 0.3544
Delay 30 days: p-value = 0.3117
Delay 60 days: p-value = 0.0019
Delay 90 days: p-value = 0.8396
Delay 180 days: p-value = 0.9573
Delay 365 days: p-value = 0.9954
Airline: American
Delay 7 days: p-value = 0.0081
Delay 14 days: p-value = 0.4480
Delay 30 days: p-value = 0.6634
Delay 60 days: p-value = 0.0010
Delay 90 days: p-value = 0.8553
Delay 180 days: p-value = 0.8069
Delay 365 days: p-value = 0.7278
Airline: Allegiant
Delay 7 days: p-value = 0.1628
Delay 14 days: p-value = 0.0006
Delay 30 days: p-value = 0.0307
Delay 60 days: p-value = 0.1716
Delay 90 days: p-value = 0.2446
Delay 180 days: p-value = 0.3790
Delay 365 days: p-value = 0.6916
Airline: JetBlue
Delay 7 days: p-value = 0.0041
Delay 14 days: p-value = 0.0885
Delay 30 days: p-value = 0.3871
Delay 60 days: p-value = 0.0087
Delay 90 days: p-value = 0.8219
Delay 180 days: p-value = 0.8590
Delay 365 days: p-value = 0.8681
Airline: Hawaiian
Delay 7 days: p-value = 0.0063
Delay 14 days: p-value = 0.1649
Delay 30 days: p-value = 0.0629
Delay 60 days: p-value = 0.0354
Delay 90 days: p-value = 0.8775
Delay 180 days: p-value = 0.0840
Delay 365 days: p-value = 0.9624
Airline: Southwest
Delay 7 days: p-value = 0.0503
Delay 14 days: p-value = 0.3359
Delay 30 days: p-value = 0.0986
Delay 60 days: p-value = 0.0359
Delay 90 days: p-value = 0.7465
Delay 180 days: p-value = 0.7389
Delay 365 days: p-value = 0.9793
Airline: Spirit
Delay 7 days: p-value = 0.1685
Delay 14 days: p-value = 0.5631
Delay 30 days: p-value = 0.0255
Delay 60 days: p-value = 0.0064
Delay 90 days: p-value = 0.7538
Delay 180 days: p-value = 0.9444
Delay 365 days: p-value = 0.8219
Airline: Delta
Delay 7 days: p-value = 0.0015
Delay 14 days: p-value = 0.8427
Delay 30 days: p-value = 0.4765
Delay 60 days: p-value = 0.0051
Delay 90 days: p-value = 0.7383
Delay 180 days: p-value = 0.6960
Delay 365 days: p-value = 0.8663
Airline: JETS
Delay 7 days: p-value = 0.0043
Delay 14 days: p-value = 0.2045
Delay 30 days: p-value = 0.1287
Delay 60 days: p-value = 0.0041
Delay 90 days: p-value = 0.7339
Delay 180 days: p-value = 0.8117
Delay 365 days: p-value = 0.8702

# create combined line chart for all airlines with significance stars for correlations
p_values_data = {
    'United': [0.0221, 0.1515, 0.2333, 0.0002, 0.5484, 0.4297, 0.8208],
    'Alaskan': [0.0558, 0.3544, 0.3117, 0.0019, 0.8396, 0.9573, 0.9954],
    'American': [0.0081, 0.4480, 0.6634, 0.0010, 0.8553, 0.8069, 0.7278],
    'Allegiant': [0.1628, 0.0006, 0.0307, 0.1716, 0.2446, 0.3790, 0.6916],
    'JetBlue': [0.0041, 0.0885, 0.3871, 0.0087, 0.8219, 0.8590, 0.8681],
    'Hawaiian': [0.0063, 0.1649, 0.0629, 0.0354, 0.8775, 0.0840, 0.9624],
    'Southwest': [0.0503, 0.3359, 0.0986, 0.0359, 0.7465, 0.7389, 0.9793],
    'Spirit': [0.1685, 0.5631, 0.0255, 0.0064, 0.7538, 0.9444, 0.8219],
    'Delta': [0.0015, 0.8427, 0.4765, 0.0051, 0.7383, 0.6960, 0.8663],
    'JETS': [0.0043, 0.2045, 0.1287, 0.0041, 0.7339, 0.8117, 0.8702]
}

# Initialize the plot
plt.figure(figsize=(14, 10))

# Plotting the correlation values for each airline
for airline, corr_values in correlations.items():
    p_values = p_values_data[airline]
    plt.plot(delays, corr_values, marker='o', label=airline)
    # Adding stars for significant correlations
    for delay, p_value, corr in zip(delays, p_values, corr_values):
        if p_value < 0.05:
            plt.plot(delay, corr, 'r*', markersize=12)

plt.title('Correlation of Oil Returns with Airline Returns Over Various Time Delays')
plt.xlabel('Time Delay (days)')
plt.ylabel('Correlation Coefficient')
plt.legend()
plt.grid(False)
plt.show()

#filter by carriers we are analyzing
financials = financials[financials['CARRIER'].isin(['AA', 'DL', 'WN', 'UA', 'AS', 'F9', 'HA', 'G4', 'B6', 'NK'])]
financials

grouped_income = pd.DataFrame()
grouped_income = financials.groupby(['UNIQUE_CARRIER', 'YEAR', 'QUARTER'])['OP_PROFIT_LOSS'].sum()
grouped_income_df = grouped_income.reset_index()

grouped_income_new = pd.DataFrame()
grouped_income_new = financials.groupby(['UNIQUE_CARRIER', 'YEAR', 'QUARTER'])['OP_PROFIT_LOSS'].sum()
grouped_income_df_new = grouped_income.reset_index()

# set back quarter by 1 since these operating losses aren't known until the future
# we want to use current operating losses (posted at end of quarter) to predict declines
grouped_income_df_new['QUARTER'] = grouped_income_df_new['QUARTER'] - 1

for i in grouped_income_df_new.index:
    if grouped_income_df_new.at[i, 'QUARTER'] == 0:
        grouped_income_df_new.at[i, 'QUARTER'] = 4
        grouped_income_df_new.at[i, 'YEAR'] -= 1


grouped_income_df_new

plt.figure(figsize=(10, 6))
for carrier in grouped_income_df['UNIQUE_CARRIER'].unique():
    subset = grouped_income_df[grouped_income_df['UNIQUE_CARRIER'] == carrier]
    plt.plot(subset['YEAR'].astype(str) + ' Q' + subset['QUARTER'].astype(str),
             subset['OP_PROFIT_LOSS'], label=carrier, marker='o')

plt.title('Operational Profit/Loss by Quarter for Market Carriers (2020-2023)')
plt.xlabel('Year and Quarter')
plt.ylabel('Operational Profit/Loss')
plt.xticks(rotation=45)
plt.legend()
plt.grid(False)
plt.tight_layout()
plt.show()

for airline, df in financ_airlines.items():
    # calculate difference in two day consecutive 'Close' price
    df['Price Change'] = df['Close'].diff()

    # label if price went up or down
    df['Trend'] = df['Price Change'].apply(lambda x: 'up' if x > 0 else ('down' if x < 0 else 'no change'))

def calculate_transition_probabilities(df):
    # shifted column to compare each day's trend with the previous day's
    df['Prev Trend'] = df['Trend'].shift(1)

    # transition state
    transitions = ['up-up', 'down-down', 'up-down', 'down-up']

    # initialize dictionary to hold the counts of each transition
    transition_counts = {transition: 0 for transition in transitions}

    # loop through and count the transitions
    for prev, current in zip(df['Prev Trend'][1:], df['Trend'][1:]):  # Skip the first entry (NaN)
        if prev == 'up' and current == 'up':
            transition_counts['up-up'] += 1
        elif prev == 'down' and current == 'down':
            transition_counts['down-down'] += 1
        elif prev == 'up' and current == 'down':
            transition_counts['up-down'] += 1
        elif prev == 'down' and current == 'up':
            transition_counts['down-up'] += 1

    # sum the total number of transitions and calculate probabilities
    total_transitions = sum(transition_counts.values())
    transition_probabilities = {transition: count / total_transitions for transition, count in transition_counts.items()}
    return transition_probabilities


transitions = ['up-up', 'up-down', 'down-up', 'down-down']

transition_matrix = pd.DataFrame(columns=transitions)

for airline, df in financ_airlines.items():
    transition_probs = calculate_transition_probabilities(df)
    transition_matrix.loc[airline] = transition_probs.values()

print(transition_matrix)

              up-up   up-down   down-up  down-down
United     0.266790  0.239463  0.247337   0.246410
Alaskan    0.250348  0.238737  0.255922   0.254993
American   0.250117  0.264610  0.242637   0.242637
Allegiant  0.236793  0.264133  0.249305   0.249768
JetBlue    0.225991  0.255614  0.258958   0.259436
Hawaiian   0.233882  0.252235  0.257412   0.256471
Southwest  0.252666  0.242930  0.252202   0.252202
Spirit     0.215377  0.263085  0.261232   0.260306
Delta      0.260159  0.231667  0.255021   0.253153
Frontier   0.203927  0.324773  0.234139   0.237160
JETS       0.264720  0.235045  0.249647   0.250589

plt.figure(figsize=(10, 8))
plt.title('Transition Probability Heatmap')

sns.heatmap(transition_matrix, annot=True, cmap='Blues', fmt=".3f")

plt.title('Transition Probability Heatmap')
plt.ylabel('Airlines')
plt.xlabel('Transition Type')

plt.show()

# Load in flight Data
flight_cols = ['MKT_UNIQUE_CARRIER', 'OP_UNIQUE_CARRIER','YEAR', 'QUARTER', 'MONTH', 'DAY_OF_MONTH',
       'DAY_OF_WEEK', 'FL_DATE', 'MKT_CARRIER','OP_CARRIER', 'TAIL_NUM',
        'ORIGIN', 'DEST','CRS_DEP_TIME', 'DEP_TIME', 'DEP_DELAY',
        'DEP_DELAY_NEW', 'DEP_DEL15','DEP_DELAY_GROUP',
        'TAXI_OUT', 'WHEELS_OFF','WHEELS_ON', 'TAXI_IN', 'CRS_ARR_TIME',
        'ARR_TIME', 'ARR_DELAY','ARR_DELAY_NEW', 'ARR_DEL15', 'ARR_DELAY_GROUP',
        'CANCELLED', 'CANCELLATION_CODE', 'DIVERTED']

airline_code = ['AA', 'DL', 'WN', 'UA', 'AS', 'F9', 'HA', 'G4', 'B6', 'NK']
file_dict = {}
for airline in airline_code:
  file_dict[airline] = glob.glob('../cmps3160-project/202*-*/{}*.csv'.format(airline))

dtype_cols = {'MKT_UNIQUE_CARRIER':'str', 'OP_UNIQUE_CARRIER':'str','FL_DATE':'str', 'MKT_CARRIER':'str','OP_CARRIER':'str', 'TAIL_NUM':'str',
        'ORIGIN':'str', 'DEST':'str','CANCELLED':'int', 'CANCELLATION_CODE':'str', 'DIVERTED':'int'}

def df_airline(input):
  frame  = pd.read_csv(file_dict[input][0], usecols = flight_cols, dtype = dtype_cols)
  for file in file_dict[input][1:]:
    frame2 = pd.read_csv(file, usecols = flight_cols, dtype = dtype_cols)
    frame = pd.concat([frame, frame2], ignore_index=True)
  frame['FL_DATE'] = frame['FL_DATE'].apply(lambda x: x[:-12])
  frame['FL_DATE'] = pd.to_datetime(frame['FL_DATE'])
  frame = pd.get_dummies(frame, columns=['CANCELLATION_CODE'])
  can_cols_names = {'CANCELLATION_CODE_A':'CANCEL_CODE_CARRIER', 'CANCELLATION_CODE_B':'CANCEL_CODE_WEATHER', 'CANCELLATION_CODE_C':'CANCEL_CODE_NAS', 'CANCELLATION_CODE_D':'CANCEL_CODE_SECURITY'}
  frame.rename(columns=can_cols_names, inplace=True)
  return frame

df_AA = df_airline('AA') # american
df_DL = df_airline('DL') # delta
df_WN = df_airline('WN') # southwest
df_UA = df_airline('UA') # united
df_AS = df_airline('AS') # alaskan
df_NK = df_airline('NK') # spirit
df_HA = df_airline('HA') # hawaiian
df_G4 = df_airline('G4') # allegiant
df_B6 = df_airline('B6') # jetblue
df_F9 = df_airline('F9') # frontier

flight_airlines = {'AA':df_AA, 'DL':df_DL, 'WN':df_WN,
                        'UA':df_UA, 'AS':df_AS, 'NK': df_NK,
                        'HA':df_HA, 'G4':df_G4, 'B6':df_B6,
                        'F9':df_F9}

flight_data = pd.concat([df_AA, df_DL, df_WN, df_UA, df_AS, df_NK, df_HA, df_G4, df_B6, df_F9], axis=0)

flight_data.head()

# Calculate number of delays using the binary column DEP_DEL15 or ARR_DEL15
def delays(DataFrame, airline, column):
  df = DataFrame.groupby('FL_DATE')[column].sum()
  df.rename('{a}_NUM_{b}'.format(a=airline, b=column), inplace=True)
  return df

# Calculates the number of flight per day and number of cancelations per day
def num_flights(DataFrame, airline):
  cancelled = DataFrame.groupby('CANCELLED')
  df_FL = cancelled.get_group(0)['FL_DATE'].value_counts()
  df_FL.rename('{}_NUM_FL'.format(airline), inplace=True)

  df_CAN = cancelled.get_group(1)['FL_DATE'].value_counts()
  df_CAN.rename('{}_NUM_CAN'.format(airline), inplace=True)
  return df_FL, df_CAN

# Calculates the different types of cancelations
def cancel_type(DataFrame, airline):
  cancelled = DataFrame.groupby('CANCELLED')
  df_CAN = cancelled.get_group(1)
  df_CAN = df_CAN.drop(['MKT_UNIQUE_CARRIER', 'YEAR', 'QUARTER', 'MONTH', 'DAY_OF_MONTH',
       'DAY_OF_WEEK', 'MKT_CARRIER', 'OP_UNIQUE_CARRIER',
       'OP_CARRIER', 'TAIL_NUM', 'ORIGIN', 'DEST', 'CRS_DEP_TIME', 'DEP_TIME',
       'DEP_DELAY', 'DEP_DELAY_NEW', 'DEP_DEL15', 'DEP_DELAY_GROUP',
       'TAXI_OUT', 'WHEELS_OFF', 'WHEELS_ON', 'TAXI_IN', 'CRS_ARR_TIME',
       'ARR_TIME', 'ARR_DELAY', 'ARR_DELAY_NEW', 'ARR_DEL15',
       'ARR_DELAY_GROUP', 'CANCELLED', 'DIVERTED'], axis=1)
  df_CAN = df_CAN.groupby('FL_DATE').sum()
  return df_CAN

# Number of Departure Delays
df_DEP_DEL = delays(df_AA, 'AA', 'DEP_DEL15')

# Number of Arrival Delays
df_ARR_DEL = delays(df_AA, 'AA', 'ARR_DEL15')

# DataFrames for Number of Flights and Number of Cancelations
df_NUM_FL, df_NUM_CAN = num_flights(df_AA, 'AA')

# Proportion of ALL flights that are canceled
df_PROP_CAN_FL = pd.DataFrame()
df_PROP_CAN_FL['AA_PROP_CAN_FL'] = df_NUM_CAN/df_NUM_FL


for airline in airline_code[1:]:
  df_DEP_DEL = pd.concat([df_DEP_DEL, delays(flight_airlines[airline], airline, 'DEP_DEL15')], axis=1)
  df_ARR_DEL = pd.concat([df_ARR_DEL, delays(flight_airlines[airline], airline,'ARR_DEL15')], axis=1)

  num_fl, num_can = num_flights(flight_airlines[airline], airline)
  prop = pd.DataFrame()
  prop['{}_PROP_CAN_FL'.format(airline)] = num_can / num_fl
  df_NUM_FL = pd.concat([df_NUM_FL, num_fl], axis=1)
  df_NUM_CAN = pd.concat([df_NUM_CAN, num_can], axis=1)
  df_PROP_CAN_FL = pd.concat([df_PROP_CAN_FL, prop], axis=1)

df_DEP_DEL = df_DEP_DEL.sort_index()
df_ARR_DEL = df_ARR_DEL.sort_index()
df_NUM_FL = df_NUM_FL.sort_index()
df_NUM_CAN = df_NUM_CAN.sort_index()
df_PROP_CAN_FL =  df_PROP_CAN_FL.sort_index()

# Plotting Total Daily Depature Delays
fig, ax = plt.subplots(figsize=(14,7))

bottom = np.zeros(len(df_ARR_DEL))
for col in df_DEP_DEL.columns:
  p = ax.bar(df_DEP_DEL.index.strftime('%Y-%m-%d'), df_DEP_DEL[col], label = col[0:2], bottom = bottom)
  bottom += df_DEP_DEL[col]

plt.title('Total Daily Depature Delays')
plt.xlabel('Date')
plt.ylabel('Number of Departure Delays')
ax.set_xticks(ax.get_xticks()[::7])
plt.xticks(rotation=90)
ax.legend(title = 'Market Carrier', loc='upper right')
plt.show()

# Plotting Total Daily Arrival Delays
fig, ax = plt.subplots(figsize=(14,7))

bottom = np.zeros(len(df_ARR_DEL))
for col in df_ARR_DEL.columns:
  p = ax.bar(df_ARR_DEL.index.strftime('%Y-%m-%d'), df_ARR_DEL[col], label = col[0:2], bottom = bottom)
  bottom += df_ARR_DEL[col]

plt.title('Total Arrival Delays')
plt.xlabel('Date')
plt.ylabel('Number of Arrival Delays')
ax.set_xticks(ax.get_xticks()[::7])
plt.xticks(rotation=90)
ax.legend(title = 'Market Carrier', loc='upper right')
plt.show()

# Plotting Market Carrier Daily Depature Delays
fig, ax = plt.subplots(figsize=(14,7))
legend_DEP_DEL = {}
for col in df_DEP_DEL.columns:
  plt.plot(df_DEP_DEL.index.strftime('%Y-%m-%d'), df_DEP_DEL[col])
  legend_DEP_DEL[col[0:2]] = col

plt.title('Market Carrier Daily Depature Delays')
plt.xlabel('Date')
plt.ylabel('Number of Depature Delays')
ax.set_xticks(ax.get_xticks()[::7])
plt.xticks(rotation=90)
ax.legend(legend_DEP_DEL, title = 'Market Carrier', loc='upper right')
plt.show()

#Plotting Market Carrier Daily Number of Flights

fig, ax = plt.subplots(figsize=(14,7))
legend_NUM_FL= {}
for col in df_NUM_FL.columns:
  plt.plot(df_NUM_FL.index.strftime('%Y-%m-%d'), df_NUM_FL[col])
  legend_NUM_FL[col[0:2]] = col

plt.title('Market Carrrier Daily Number of Flights')
plt.xlabel('Date')
plt.ylabel('Number of Flights')
ax.set_xticks(ax.get_xticks()[::7])
plt.xticks(rotation=90)
ax.legend(legend_NUM_FL, title = 'Market Carrier', loc='upper right')
plt.show()

#Plotting Market Carrier Daily Number of Cancelations

fig, ax = plt.subplots(figsize=(14,7))
legend_NUM_CAN= {}
for col in df_NUM_CAN.columns:
  plt.plot(df_NUM_CAN.index.strftime('%Y-%m-%d'), df_NUM_CAN[col])
  legend_NUM_CAN[col[0:2]] = col

plt.title('Market Carrier Daily Number of Canelations')
plt.xlabel('Date')
plt.ylabel('Number of Canelations')
ax.set_xticks(ax.get_xticks()[::7])
plt.xticks(rotation=90)
ax.legend(legend_NUM_CAN, title = 'Market Carrier', loc='upper right')
plt.show()

# Create a Correlation Martix for the Proportion of Cancelled Flights by Market Carrier
MKT_code = {'AA':'American', 'DL':'Delta', 'WN':'Southwest',
                        'UA':'United', 'AS': 'Alaskan', 'NK':'Spirit',
                        'HA':'Hawaiian', 'G4':'Allegiant', 'B6':'JetBlue',
                        'F9':'Frontier'}

prop_can_corr_matrix = df_PROP_CAN_FL.corr()

label_names = []

for item in list(prop_can_corr_matrix.columns):
  code = item[0:2]
  name = MKT_code[code]
  label_names.append(name)

cmap = sns.diverging_palette(10, 130, as_cmap=True)
plt.figure(figsize=(10, 8))
sns.heatmap(prop_can_corr_matrix, annot=True, cmap = 'Blues', vmin=0, vmax=1,
            xticklabels=label_names,
            yticklabels=label_names)
plt.title('Correlation Matrix of the Proportion of Cancelled Flights (by Market Carrier)')
plt.show()

def inner_op_mkt(mkt_carrier, airline):
  df = mkt_carrier[airline].to_frame().rename(columns={'FL_DATE':'NUM_SCH_FL'}).reset_index()
  df = df.set_index('FL_DATE')

  op_not_mkt = df[df['OP_UNIQUE_CARRIER']!= airline]
  op_not_mkt = op_not_mkt['NUM_SCH_FL'].rename('OP_NOT_MKT').to_frame().reset_index()
  op_not_mkt = op_not_mkt.groupby('FL_DATE')['OP_NOT_MKT'].sum()

  op_mkt = df[df['OP_UNIQUE_CARRIER'] == airline]
  op_mkt = op_mkt['NUM_SCH_FL'].rename('OP_MKT')

  df_mkt = op_mkt.to_frame().join(op_not_mkt, on='FL_DATE')
  df_mkt['PROP_OP_BY_MKT'] = df_mkt['OP_MKT']/(df_mkt['OP_MKT']+ df_mkt['OP_NOT_MKT'])

  df_mkt = df_mkt.drop(['OP_MKT', 'OP_NOT_MKT'], axis=1)
  return df_mkt

# status {0, 1}
# Flights = 0
# Canceled = 1
def all_op_mkt(status):
  if status == 0 or  status == 1:
    flights_status = flight_data.groupby('CANCELLED')
    fl_status = flights_status.get_group(status)

    fl_type = {0:'FL', 1:'CAN'}
  else:
    return None
  airline_code = ['AA', 'DL', 'WN', 'UA', 'AS', 'F9', 'HA', 'G4', 'B6', 'NK']

  mkt_carrier = fl_status.groupby(['MKT_UNIQUE_CARRIER', 'OP_UNIQUE_CARRIER'])['FL_DATE'].value_counts()
  df_op_mkt = inner_op_mkt(mkt_carrier, airline_code[0]).add_suffix('_{}'.format(airline_code[0]))
  df_op_mkt = df_op_mkt.add_prefix('{}_'.format(fl_type[status]))

  for airline in airline_code[1:]:
    df_mkt = inner_op_mkt(mkt_carrier, airline)

    if  df_mkt['PROP_OP_BY_MKT'].isnull().all() == True:
      df_mkt['PROP_OP_BY_MKT'] = 0
      df_mkt = df_mkt.add_suffix('_{}'.format(airline))
    else:
      df_mkt = df_mkt.add_suffix('_{}'.format(airline))

    df_mkt = df_mkt.add_prefix('{}_'.format(fl_type[status]))


    df_op_mkt = df_op_mkt.join(df_mkt, on='FL_DATE')

  df_op_mkt.dropna(how='all', axis=1, inplace=True)
  df_op_mkt.fillna(0, inplace=True)
  df_op_mkt = df_op_mkt.sort_index()


  return df_op_mkt

FL_op_mkt = all_op_mkt(0) # only flights
CAN_op_mkt = all_op_mkt(1) # only cancelled flight

# Plotting Proportion of cancelations operated by the market carrier
fig, ax = plt.subplots(figsize=(14,7))

bottom = np.zeros(len( CAN_op_mkt))
for col in CAN_op_mkt.replace(0,np.nan).dropna(axis=1,how="all").columns:
  p = ax.bar(CAN_op_mkt.index.strftime('%Y-%m-%d'), CAN_op_mkt[col], label = col[-2:], bottom = bottom)
  bottom += CAN_op_mkt[col]

plt.title('Proportion of Cancellations Operated by Market Carrier')
plt.xlabel('Date')
plt.ylabel('Proportion of Cancellations')
ax.set_xticks(ax.get_xticks()[::7])
plt.xticks(rotation=90)
ax.legend(title = 'Market Carrier', loc='upper right')
plt.show()

# Load in T-100 Domestic Segment (U.S. Carriers)
T_100_lst = glob.glob('../cmps3160-project/T-100 Domestic Segment (U.S. Carriers)/*.csv')

df_air_stats = pd.read_csv(T_100_lst[0])

for item in T_100_lst[1:]:
  frame =  pd.read_csv(item)
  df_air_stats = pd.concat([df_air_stats, frame], ignore_index=True)

df_air_stats.head()

df_stat_groups = df_air_stats.groupby('UNIQUE_CARRIER')

full_seats={}
freight_weight={}

# Proportion of Full Seats
for airline in airline_code:
  inner_dict = {}
  yr_month_group = df_air_stats.groupby(['YEAR', 'MONTH'])
  for i in yr_month_group.groups.keys():
    frame = yr_month_group.get_group(i)
    yr = frame['YEAR'].unique().tolist()[0]
    month = frame['MONTH'].unique().tolist()[0]
    month_prop_full = frame['PASSENGERS'].sum() / frame['SEATS'].sum()
    inner_dict[i] = month_prop_full
  full_seats[airline] = inner_dict

# Proportion of Weight from Freight and Mail
for airline in airline_code:
  inner_dict = {}
  yr_month_group = df_air_stats.groupby(['YEAR', 'MONTH'])
  for i in yr_month_group.groups.keys():
    frame = yr_month_group.get_group(i)
    yr = frame['YEAR'].unique().tolist()[0]
    month = frame['MONTH'].unique().tolist()[0]
    month_prop_freight = (frame['FREIGHT'].sum() + frame['MAIL'].sum())/ frame['PAYLOAD'].sum()
    inner_dict[i] = month_prop_freight
  freight_weight[airline] = inner_dict

# DataFrames of proportion of full seats and proportion of weight from freight and mail
df_full_seats = pd.DataFrame.from_dict(full_seats).add_suffix('_PROP_FULL_SEATS')
df_freight_weight = pd.DataFrame.from_dict(freight_weight).add_suffix('_PROP_FREIGHT/MAIL')

df_full_seats.index.rename(['YEAR', 'MONTH'], inplace=True)
df_freight_weight.index.rename(['YEAR', 'MONTH'], inplace=True)

dataframe2  = pd.DataFrame()
date_range = pd.date_range(start=df_AA['FL_DATE'].min(),end=df_AA['FL_DATE'].max())
dataframe2['FL_DATE'] = date_range
dataframe2['YEAR'] = pd.DatetimeIndex(dataframe2['FL_DATE']).year
dataframe2['MONTH'] = pd.DatetimeIndex(dataframe2['FL_DATE']).month

df_prop_full_seats = pd.merge(dataframe2, df_full_seats, on=['YEAR', 'MONTH'], validate="many_to_one")
df_prop_full_seats.drop(['YEAR', 'MONTH'], axis=1, inplace=True)
df_prop_full_seats.set_index('FL_DATE', inplace=True)

df_prop_freight_weight = pd.merge(dataframe2, df_freight_weight, on=['YEAR', 'MONTH'], validate="many_to_one")
df_prop_freight_weight.drop(['YEAR', 'MONTH'], axis=1, inplace=True)
df_prop_freight_weight.set_index('FL_DATE', inplace=True)

def T100_stat_by_airline(summ_stat_frame, type, startdate, enddate):
  dataframe2  = pd.DataFrame()
  date_range = pd.date_range(start=startdate,end=enddate)
  dataframe2['FL_DATE'] = date_range
  dataframe2['YEAR'] = pd.DatetimeIndex(dataframe2['FL_DATE']).year
  dataframe2['MONTH'] = pd.DatetimeIndex(dataframe2['FL_DATE']).month

  df_mean_stats = summ_stat_frame.loc[airline].add_prefix('{a}_{b}_'.format(a=airline, b=type))
  df_mean_stats = pd.merge(dataframe2, df_mean_stats, on=['YEAR', 'MONTH'], validate="many_to_one")
  df_mean_stats.drop(['YEAR', 'MONTH'], axis=1, inplace=True)
  df_mean_stats.set_index('FL_DATE', inplace=True)
  return df_mean_stats

# Dictionary of all flight metrics DataFrames
flight_metics = {'Departure' : df_DEP_DEL, 'Arrival' : df_ARR_DEL,
                 'Fight Count' : df_NUM_FL, 'Cancel Count': df_NUM_CAN,
                 'Prop Flights Operated by Market' : FL_op_mkt,
                 'Prop Canceled Operated by Market' : CAN_op_mkt,
                 'Prop Full Seats' : df_full_seats, 'Prop Freight/Mail': df_freight_weight
                 }

carrier_group_dict = {'1':'Regional Carriers (including Large, Medium, Commuter, Small Certified)',
                      '2':'National Carriers', '3':'Major Carriers'}

T100_carrier_groups = df_air_stats.groupby(['CARRIER_GROUP', 'YEAR', 'MONTH'])

T100_mean = T100_carrier_groups['DEPARTURES_SCHEDULED', 'DEPARTURES_PERFORMED', 'PAYLOAD', 'SEATS', 'PASSENGERS',
 'FREIGHT', 'MAIL', 'DISTANCE', 'RAMP_TO_RAMP', 'AIR_TIME', 'DISTANCE_GROUP'].mean()

Major_Mean = T100_mean.loc[3, :].add_prefix('Major_Carriers_Mean_')
National_Mean = T100_mean.loc[2, :].add_prefix('National_Carriers_Mean_')
Regional_Mean = T100_mean.loc[2, :].add_prefix('Regional_Carriers_Mean_')

<ipython-input-54-3fba41ab82e6>:6: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.
  T100_mean = T100_carrier_groups['DEPARTURES_SCHEDULED', 'DEPARTURES_PERFORMED', 'PAYLOAD', 'SEATS', 'PASSENGERS',

df_air_stats.head()

df_AA_op_stat = pd.DataFrame() # american
df_DL_op_stat = pd.DataFrame() # delta
df_WN_op_stat = pd.DataFrame() # southwest
df_UA_op_stat = pd.DataFrame() # united
df_AS_op_stat = pd.DataFrame() # alaskan
df_NK_op_stat = pd.DataFrame() # spirit
df_HA_op_stat = pd.DataFrame() # hawaiian
df_G4_op_stat = pd.DataFrame() # allegiant
df_B6_op_stat = pd.DataFrame() # jetblue
df_F9_op_stat = pd.DataFrame() # frontier

airlines_operations_stats = {'AA':df_AA_op_stat, 'DL':df_DL_op_stat, 'WN':df_WN_op_stat,
                        'UA':df_UA_op_stat, 'AS':df_AS_op_stat, 'NK': df_NK_op_stat,
                        'HA':df_HA_op_stat, 'G4':df_G4_op_stat, 'B6':df_B6_op_stat,
                        'F9':df_F9_op_stat}

startdate = df_AA['FL_DATE'].min()
enddate = df_AA['FL_DATE'].max()

df_airline_stat = df_air_stats.groupby(['UNIQUE_CARRIER', 'YEAR', 'MONTH'])
df_mean_stats = df_airline_stat[['DEPARTURES_SCHEDULED', 'DEPARTURES_PERFORMED', 'PAYLOAD', 'SEATS', 'PASSENGERS','FREIGHT', 'MAIL', 'DISTANCE', 'RAMP_TO_RAMP', 'AIR_TIME']].mean()
df_std_stats = df_airline_stat[['DEPARTURES_SCHEDULED', 'DEPARTURES_PERFORMED', 'PAYLOAD', 'SEATS', 'PASSENGERS','FREIGHT', 'MAIL', 'DISTANCE', 'RAMP_TO_RAMP', 'AIR_TIME']].std()
df_median_stats = df_airline_stat[['DEPARTURES_SCHEDULED', 'DEPARTURES_PERFORMED', 'PAYLOAD', 'SEATS', 'PASSENGERS','FREIGHT', 'MAIL', 'DISTANCE', 'RAMP_TO_RAMP', 'AIR_TIME']].median()

from datetime import datetime
frames_lst = flight_metics.values()
for frame in frames_lst:
  cols = frame.columns
  #print(frame)
  if frame.index.dtype =='datetime64[ns]':
    frame.index.rename('FL_DATE', inplace=True)
    for airline in airline_code:
      air_cols = cols.str.contains(airline)
      frame2 = frame.loc[:, air_cols]
      airlines_operations_stats[airline] = pd.concat([airlines_operations_stats[airline], frame2], axis=1)
  else:
    dataframe  = pd.DataFrame()
    date_range = pd.date_range(start=startdate,end=enddate)
    dataframe['FL_DATE'] = date_range
    dataframe['YEAR'] = pd.DatetimeIndex(dataframe['FL_DATE']).year
    dataframe['MONTH'] = pd.DatetimeIndex(dataframe['FL_DATE']).month
    for airline in airline_code:
      air_cols = cols.str.contains(airline)
      frame2 = frame.loc[:, air_cols]
      frame2 = pd.merge(dataframe, frame2, on=['YEAR', 'MONTH'])
      frame2.drop(['YEAR', 'MONTH'], axis=1, inplace=True)
      frame2.set_index('FL_DATE', inplace=True)
      airlines_operations_stats[airline] = pd.concat([airlines_operations_stats[airline], frame2], axis=1)

for airline in airline_code:
    frame_can_type = cancel_type(flight_airlines[airline], airline)

    for item in ['CANCEL_CODE_CARRIER', 'CANCEL_CODE_WEATHER', 'CANCEL_CODE_NAS','CANCEL_CODE_SECURITY']:
      if item in frame_can_type.columns:
        continue
      else:
        frame_can_type[item] = 0

    frame3 = T100_stat_by_airline(df_mean_stats, 'MEAN', startdate, enddate)
    frame4 = T100_stat_by_airline(df_std_stats, 'STD', startdate, enddate)
    frame5 = T100_stat_by_airline(df_median_stats, 'MEDIAN', startdate, enddate)
    airlines_operations_stats[airline] = pd.concat([airlines_operations_stats[airline], frame_can_type,frame3, frame4, frame5], axis=1)
    airlines_operations_stats[airline]['AIRLINE'] = airline

df_american = pd.DataFrame()
df_delta = pd.DataFrame()
df_southwest = pd.DataFrame()
df_united = pd.DataFrame()
df_alaskan = pd.DataFrame()
df_spirit = pd.DataFrame()
df_hawaiian = pd.DataFrame()
df_allegiant = pd.DataFrame()
df_jetblue = pd.DataFrame()
df_jets = pd.DataFrame()
df_crude_oil = pd.DataFrame()

delay_decline = {'United': df_united, 'Alaskan': df_alaskan, 'American': df_american,
                   'Allegiant': df_allegiant, 'JetBlue': df_jetblue, 'Hawaiian': df_hawaiian,
                   'Southwest': df_southwest, 'Spirit': df_spirit, 'Delta': df_delta,
                   'JETS' : df_jets}

#crude oil data
crude_frame = financ_OIL.reset_index()
mask = (crude_frame['Date'] >= startdate) & (crude_frame['Date'] <= enddate)
crude_frame = crude_frame.loc[mask]
crude_frame.rename(columns={'Date':'DATE'}, inplace=True)
crude_frame = crude_frame.set_index('DATE')
crude_frame = crude_frame.add_prefix('Crude Oil ')

for airline in delay_decline.keys():
  #stock data
  fin_frame1 = financ_airlines[airline]
  fin_frame = fin_frame1.reset_index()
  mask = (fin_frame['Date'] >= startdate) & (fin_frame['Date'] <= enddate)
  fin_frame = fin_frame.loc[mask]
  fin_frame.rename(columns={'Date':'DATE'}, inplace=True)
  fin_frame = fin_frame.set_index('DATE')
  if airline == 'JETS':
    fin_frame = fin_frame.add_prefix('JETS ')
    all_airline = pd.DataFrame()
    for code in airline_code:
      air_frame1 = airlines_operations_stats[code]
      air_frame = air_frame1.reset_index()
      air_frame.rename(columns={'FL_DATE':'DATE'}, inplace=True)
      air_frame = air_frame.set_index('DATE')
      all_airline = pd.concat([all_airline, all_airline], axis=1)
  else:
    #airline data
    code = flight_airline_codes[airline]
    air_frame1 = airlines_operations_stats[code]
    air_frame = air_frame1.reset_index()
    air_frame.rename(columns={'FL_DATE':'DATE'}, inplace=True)
    air_frame = air_frame.set_index('DATE')
    delay_decline[airline] = pd.concat([air_frame, fin_frame, crude_frame], axis=1)

delay_decline['Delta'].head()

financials_drop_cols = ['NET_INCOME', 'MAIL',
       'TOTAL_PROPERTY', 'PROP_FREIGHT', 'PROP_BAG', 'TOTAL_CHARTER',
       'CHARTER_PAX', 'CHARTER_PROP', 'TOTAL_MISC_REV', 'RES_CANCEL_FEES',
       'MISC_OP_REV', 'PUB_SVC_REVENUE', 'TRANS_REVENUE', 'OP_REVENUES',
       'FLYING_OPS', 'MAINTENANCE', 'PAX_SERVICE', 'AIRCFT_SERVICES',
       'PROMOTION_SALES', 'GENERAL_ADMIN', 'GENERAL_SERVICES', 'DEPREC_AMORT',
       'TRANS_EXPENSES', 'OP_EXPENSES', 'INTEREST_LONG_DEBT',
       'INTEREST_EXP_OTH', 'FOREIGN_EX_GAINS', 'CAP_GAINS_PROP',
       'CAP_GAINS_OTHER', 'OTHER_INCOME_NET', 'NON_OP_INCOME',
       'INCOME_PRE_TAX', 'INCOME_TAX', 'INCOME_BEFORE_OTH', 'DISCONT_OPS',
       'EXTRA_ITEMS', 'INCOME_TAX_EXTRA', 'INCOME_TAX_CREDITS',
       'ACCTG_CHANGES', 'AIRLINE_ID', 'UNIQUE_CARRIER_NAME',
       'CARRIER', 'CARRIER_NAME', 'UNIQUE_CARRIER_ENTITY', 'REGION',
       'CARRIER_GROUP_NEW', 'CARRIER_GROUP']
fin_rep = financials.drop(financials_drop_cols, axis=1)
fin_rep_groups = fin_rep.groupby('UNIQUE_CARRIER')

grouped_income_df_new_g = grouped_income_df_new.groupby('UNIQUE_CARRIER')

dataframe2  = pd.DataFrame()
date_range = pd.date_range(start=df_AA['FL_DATE'].min(),end=df_AA['FL_DATE'].max())
dataframe2['FL_DATE'] = date_range
dataframe2['YEAR'] = pd.DatetimeIndex(dataframe2['FL_DATE']).year
dataframe2['QUARTER'] = pd.DatetimeIndex(dataframe2['FL_DATE']).quarter

all_fin_rep  = pd.DataFrame()
for group in grouped_income_df_new_g.groups.keys():
  fin_rep1 = grouped_income_df_new_g.get_group(group)
  fin_rep1 = fin_rep1.rename(columns={'OP_PROFIT_LOSS':'Prev Quarter OP_PROFIT_LOSS'})
  df_fin_rep = pd.merge(dataframe2, fin_rep1, on=['YEAR', 'QUARTER'])
  df_fin_rep.drop(['YEAR', 'QUARTER'], axis=1, inplace=True)
  df_fin_rep.set_index('FL_DATE', inplace=True)
  all_fin_rep = pd.concat([all_fin_rep, df_fin_rep], axis=0)

all_df = pd.DataFrame()
all_fin_rep_groups = all_fin_rep.groupby('UNIQUE_CARRIER')
for item in delay_decline.keys():
  if item != 'JETS':
    frame1 = all_fin_rep_groups.get_group(flight_airline_codes[item])
    frame = delay_decline[item]
    frame = frame.reset_index()
    frame.rename(columns={'DATE':'Date'}, inplace=True)
    frame = frame.set_index('Date')
    frame = frame.set_axis(['NUM_DEP_DEL15', 'NUM_ARR_DEL15', 'NUM_FL', 'NUM_CAN',
       'FL_PROP_OP_BY_MKT', 'CAN_PROP_OP_BY_MKT', 'PROP_FULL_SEATS',
       'PROP_FREIGHT/MAIL', 'CANCEL_CODE_CARRIER', 'CANCEL_CODE_WEATHER',
       'CANCEL_CODE_NAS', 'CANCEL_CODE_SECURITY', 'MEAN_DEPARTURES_SCHEDULED',
       'MEAN_DEPARTURES_PERFORMED', 'MEAN_PAYLOAD', 'MEAN_SEATS',
       'MEAN_PASSENGERS', 'MEAN_FREIGHT', 'MEAN_MAIL',
       'MEAN_DISTANCE', 'MEAN_RAMP_TO_RAMP', 'MEAN_AIR_TIME',
       'STD_DEPARTURES_SCHEDULED', 'STD_DEPARTURES_PERFORMED',
       'STD_PAYLOAD', 'STD_SEATS', 'STD_PASSENGERS', 'STD_FREIGHT',
       'STD_MAIL', 'STD_DISTANCE', 'STD_RAMP_TO_RAMP',
       'STD_AIR_TIME', 'MEDIAN_DEPARTURES_SCHEDULED',
       'MEDIAN_DEPARTURES_PERFORMED', 'MEDIAN_PAYLOAD',
       'MEDIAN_SEATS', 'MEDIAN_PASSENGERS', 'MEDIAN_FREIGHT',
       'MEDIAN_MAIL', 'MEDIAN_DISTANCE', 'MEDIAN_RAMP_TO_RAMP',
       'MEDIAN_AIR_TIME', 'AIRLINE', 'Open', 'High', 'Low', 'Close',
       'Adj Close', 'Volume', 'Daily Return',
        'Price Change', 'Trend','Prev Trend',
        'Crude Oil Open', 'Crude Oil High', 'Crude Oil Low',
       'Crude Oil Close', 'Crude Oil Adj Close', 'Crude Oil Volume',
       'Crude Oil Daily Return'], axis=1)
    frame_comb = pd.concat([frame, frame1], axis=1)
    frame_comb = frame_comb.reset_index()
    frame_comb = frame_comb.rename(columns={'index':'Date'})
    all_df = pd.concat([all_df, frame_comb],ignore_index=True, axis=0)
  else:
    continue
all_df = all_df.drop('UNIQUE_CARRIER', axis=1)
all_df['Trend'] =all_df['Trend'].map({'up':1, 'down':0})
all_df['Prev Trend'] =all_df['Prev Trend'].map({'up':1, 'down':0})

all_df.head()

all_df2 = all_df
all_df2 = all_df2.dropna()

shift_cols = ['Open', 'High',
       'Low', 'Close', 'Adj Close', 'Volume', 'Daily Return', 'Price Change',
       'Crude Oil Open', 'Crude Oil High',
       'Crude Oil Low', 'Crude Oil Close', 'Crude Oil Adj Close',
       'Crude Oil Volume', 'Crude Oil Daily Return']
for item in shift_cols:
  prev = 'Prev {}'.format(item)
  all_df2[prev] = all_df2[item].shift(-1)

all_df2 = all_df2.drop(shift_cols, axis=1)

<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)
<ipython-input-64-4f8ca05e1355>:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  all_df2[prev] = all_df2[item].shift(-1)

all_df2.columns

data = all_df2.drop('Date', axis=1)
data = data.dropna()

data3 = data.drop(['MEAN_DEPARTURES_PERFORMED', 'MEAN_PAYLOAD', 'MEAN_SEATS',
       'MEAN_PASSENGERS', 'MEAN_FREIGHT', 'MEAN_MAIL', 'MEAN_DISTANCE',
       'MEAN_RAMP_TO_RAMP', 'MEAN_AIR_TIME', 'STD_DEPARTURES_SCHEDULED',
       'STD_DEPARTURES_PERFORMED', 'STD_PAYLOAD', 'STD_SEATS',
       'STD_PASSENGERS', 'STD_FREIGHT', 'STD_MAIL', 'STD_DISTANCE',
       'STD_RAMP_TO_RAMP', 'STD_AIR_TIME', 'MEDIAN_DEPARTURES_SCHEDULED',
       'MEDIAN_DEPARTURES_PERFORMED', 'MEDIAN_PAYLOAD', 'MEDIAN_SEATS',
       'MEDIAN_PASSENGERS', 'MEDIAN_FREIGHT', 'MEDIAN_MAIL', 'MEDIAN_DISTANCE',
       'MEDIAN_RAMP_TO_RAMP', 'MEDIAN_AIR_TIME'], axis=1)

data_groups = data3.groupby('AIRLINE')
X = data_groups.get_group('HA').drop(['AIRLINE', 'Trend'], axis=1)
y = data_groups.get_group('HA')['Trend']

from sklearn.preprocessing import StandardScaler
from sklearn.manifold import TSNE


TSNE_fin_data = all_df.drop(['Date','NUM_DEP_DEL15', 'NUM_ARR_DEL15', 'NUM_FL', 'NUM_CAN',
       'FL_PROP_OP_BY_MKT', 'CAN_PROP_OP_BY_MKT', 'PROP_FULL_SEATS',
       'PROP_FREIGHT/MAIL', 'CANCEL_CODE_CARRIER', 'CANCEL_CODE_WEATHER',
       'CANCEL_CODE_NAS', 'CANCEL_CODE_SECURITY', 'MEAN_DEPARTURES_SCHEDULED',
       'MEAN_DEPARTURES_PERFORMED', 'MEAN_PAYLOAD', 'MEAN_SEATS',
       'MEAN_PASSENGERS', 'MEAN_FREIGHT', 'MEAN_MAIL',
       'MEAN_DISTANCE', 'MEAN_RAMP_TO_RAMP', 'MEAN_AIR_TIME',
       'STD_DEPARTURES_SCHEDULED', 'STD_DEPARTURES_PERFORMED',
       'STD_PAYLOAD', 'STD_SEATS', 'STD_PASSENGERS', 'STD_FREIGHT',
       'STD_MAIL', 'STD_DISTANCE', 'STD_RAMP_TO_RAMP',
       'STD_AIR_TIME', 'MEDIAN_DEPARTURES_SCHEDULED',
       'MEDIAN_DEPARTURES_PERFORMED', 'MEDIAN_PAYLOAD',
       'MEDIAN_SEATS', 'MEDIAN_PASSENGERS', 'MEDIAN_FREIGHT',
       'MEDIAN_MAIL', 'MEDIAN_DISTANCE', 'MEDIAN_RAMP_TO_RAMP',
       'MEDIAN_AIR_TIME'], axis=1)
TSNE_fin_data = TSNE_fin_data.dropna()

X_T = TSNE_fin_data.drop('AIRLINE', axis=1)
y_T = TSNE_fin_data['AIRLINE']
X_T_std = StandardScaler().fit_transform(X_T)
n_components = 2
tsne = TSNE(n_components)
X_tsne = tsne.fit_transform(X_T_std)
X_tsne_data = np.vstack((X_tsne.T, y_T)).T
df_tsne = pd.DataFrame(X_tsne_data, columns=['Dim1', 'Dim2', 'AIRLINE'])

plt.figure(figsize=(8, 8))
sns.scatterplot(data=df_tsne, hue='AIRLINE', x='Dim1', y='Dim2').set(title='TSNE Stock and Financial Data')
plt.show()

TSNE_data = all_df.drop('Date', axis=1)
TSNE_data = TSNE_data.dropna()

X_T = TSNE_data.drop('AIRLINE', axis=1)
y_T = TSNE_data['AIRLINE']
X_T_std = StandardScaler().fit_transform(X_T)
n_components = 2
tsne = TSNE(n_components)
X_tsne = tsne.fit_transform(X_T_std)
X_tsne_data = np.vstack((X_tsne.T, y_T)).T
df_tsne = pd.DataFrame(X_tsne_data, columns=['Dim1', 'Dim2', 'AIRLINE'])

plt.figure(figsize=(8, 8))
sns.scatterplot(data=df_tsne, hue='AIRLINE', x='Dim1', y='Dim2').set(title='TSNE Flight, Stock, and Financial Data')
plt.show()

all_df_airline_groups = all_df.groupby('AIRLINE')

from sklearn.model_selection import train_test_split

def train_test_split_airline(airline, rs):
  frame = all_df_airline_groups.get_group(airline)
  frame = frame.dropna()
  X = frame.drop(['Trend','Date', 'AIRLINE'], axis=1)
  y = frame['Trend']
  X_std = StandardScaler().fit_transform(X)

  # using the train test split function
  X_std_train, X_std_test, y_train, y_test = train_test_split(X_std,y , random_state=rs,test_size=0.25, shuffle=True)
  return X_std_train, X_std_test, y_train, y_test

from sklearn.linear_model import LogisticRegression
from sklearn.metrics import mean_squared_error

df_logistic = pd.DataFrame()

for rs in [42, 104, 35]:
  air_lst = []
  logistic_acc = []
  logistic_mse = []
  for airline in all_df_airline_groups.groups.keys():
    air_lst.append(airline)

    X_std_train, X_std_test, y_train, y_test = train_test_split_airline(airline, rs)

    # all parameters not specified are set to their defaults
    logisticRegr = LogisticRegression()

    # fit model on training data
    logisticRegr.fit(X_std_train, y_train)

    # predict the testing data
    y_pred = logisticRegr.predict(X_std_test)

    # accuracy
    score = logisticRegr.score(X_std_test, y_test)
    logistic_acc.append(score)

    # MSE
    mse = mean_squared_error(y_test, y_pred)
    logistic_mse.append(mse)

  RS_col = 'Random State {}'.format(rs)
  df_l = pd.DataFrame({'Airline':air_lst, 'Accuracy':logistic_acc, 'MSE':logistic_mse})
  df_l = df_l.set_index('Airline')
  l_cols=[(RS_col,'Accuracy'),(RS_col,'MSE')]
  df_l.columns=pd.MultiIndex.from_tuples(l_cols)

  df_logistic = pd.concat([df_logistic, df_l], axis=1)
df_logistic

def train_test_split_trend(airline, rs):
  frame = data_groups.get_group(airline)
  frame = frame.dropna()
  X = frame.drop(['Trend', 'AIRLINE'], axis=1)
  y = frame['Trend']
  X_std = StandardScaler().fit_transform(X)

  # using the train test split function
  X_std_train, X_std_test, y_train, y_test = train_test_split(X_std,y , random_state=rs,test_size=0.25, shuffle=True)
  return X_std_train, X_std_test, y_train, y_test

df_logistic = pd.DataFrame()

for rs in [42, 104, 35]:
  air_lst = []
  logistic_acc = []
  logistic_mse = []
  for airline in data_groups.groups.keys():
    air_lst.append(airline)

    X_std_train, X_std_test, y_train, y_test = train_test_split_trend(airline, rs)

    # all parameters not specified are set to their defaults
    logisticRegr = LogisticRegression()

    # fit model on training data
    logisticRegr.fit(X_std_train, y_train)

    # Predict the testing data
    y_pred = logisticRegr.predict(X_std_test)

    # Accuracy
    score = logisticRegr.score(X_std_test, y_test)
    logistic_acc.append(score)

    # MSE
    mse = mean_squared_error(y_test, y_pred)
    logistic_mse.append(mse)

  RS_col = 'Random State {}'.format(rs)
  df_l = pd.DataFrame({'Airline':air_lst, 'Accuracy':logistic_acc, 'MSE':logistic_mse})
  df_l = df_l.set_index('Airline')
  l_cols=[(RS_col,'Accuracy'),(RS_col,'MSE')]
  df_l.columns=pd.MultiIndex.from_tuples(l_cols)

  df_logistic = pd.concat([df_logistic, df_l], axis=1)
df_logistic

lstm_g = all_df.groupby('AIRLINE')

# prepare data for lstm
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import MinMaxScaler
from sklearn.preprocessing import StandardScaler

# convert series to supervised learning
def series_to_supervised(data, n_in=1, n_out=1, dropnan=True):
	n_vars = 1 if type(data) is list else data.shape[1]
	df = pd.DataFrame(data)
	cols, names = list(), list()
	# input sequence (t-n, ... t-1)
	for i in range(n_in, 0, -1):
		cols.append(df.shift(i))
		names += [('var%d(t-%d)' % (j+1, i)) for j in range(n_vars)]
	# forecast sequence (t, t+1, ... t+n)
	for i in range(0, n_out):
		cols.append(df.shift(-i))
		if i == 0:
			names += [('var%d(t)' % (j+1)) for j in range(n_vars)]
		else:
			names += [('var%d(t+%d)' % (j+1, i)) for j in range(n_vars)]
	# put it all together
	agg = pd.concat(cols, axis=1)
	agg.columns = names
	# drop rows with NaN values
	if dropnan:
		agg.dropna(inplace=True)
	return agg

from sklearn.preprocessing import MinMaxScaler
def preprocessing_lstm(airline):
  # load dataset
  dataset = lstm_g.get_group(airline)
  dataset = dataset.set_index('Date')
  dataset = dataset.drop('AIRLINE', axis=1)
  values = dataset.values

  # ensure all data is float
  values = values.astype('float32')

  # train the normalization
  scaler = MinMaxScaler()
  scaler = scaler.fit(values)

  # normalize the dataset
  normalized = scaler.transform(values)

  # frame as supervised learning
  reframed = series_to_supervised(normalized, 1, 1)

  # drop columns we don't want to predict
  reframed.drop(reframed.columns[[9,10,11,12,13,14,15]], axis=1, inplace=True)

  # split into train and test sets
  values = reframed.values
  n_train_days = int(values.shape[0] // 2.5)
  train = values[:n_train_days, :]
  test = values[n_train_days:, :]

  # split into input and outputs
  train_X, train_y = train[:, :-1], train[:, -1]
  test_X, test_y = test[:, :-1], test[:, -1]

  # reshape input to be 3D [samples, timesteps, features]
  train_X = train_X.reshape((train_X.shape[0], 1, train_X.shape[1]))
  test_X = test_X.reshape((test_X.shape[0], 1, test_X.shape[1]))

  return test_X, test_y, train_X, train_y

from matplotlib import pyplot

def lstm_plot(yhat, history, test_X, test_y, airline, loss_type):
  # plot history
  pyplot.plot(history.history['loss'], label='train')
  pyplot.plot(history.history['val_loss'], label='test')
  pltitle = '{}: Train and Test Loss from the Multivariate LSTM During Training'.format(airline)
  pyplot.title(pltitle)
  pyplot.xlabel('Epoch')
  ylab = 'Train and Test Loss ({})'.format(loss_type)
  pyplot.ylabel(ylab)
  pyplot.legend()
  pyplot.show()

  test_X = test_X.reshape((test_X.shape[0], test_X.shape[2]))

  # yhat
  yhat = np.concatenate((yhat, test_X[:, 1:]), axis=1)
  yhat = yhat[:,0]

  # actual
  test_y = test_y.reshape((len(test_y), 1))
  y = np.concatenate((test_y, test_X[:, 1:]), axis=1)
  y = y[:,0]

  # calculate RMSE
  rmse = np.sqrt(mean_squared_error(y, yhat))
  print('Test RMSE: %.3f' % rmse)
  return rmse

from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM

def lstm_model(train_X, train_y, test_X, test_y, n_epochs, n_batch, loss_type):
  # design network
  model = Sequential()
  model.add(LSTM(n_epochs, input_shape=(train_X.shape[1], train_X.shape[2])))
  model.add(Dense(1))
  model.compile(loss=loss_type, optimizer='adam')
  # fit network
  history = model.fit(train_X, train_y, epochs=n_epochs, batch_size=n_batch, validation_data=(test_X, test_y), verbose=2, shuffle=False)
  # make a prediction
  yhat = model.predict(test_X)

  return yhat, history

def lstm_run(airline, n_epochs, n_batch, loss_type):
  test_X, test_y, train_X, train_y = preprocessing_lstm(airline)
  yhat, history = lstm_model(train_X, train_y, test_X, test_y, n_epochs, n_batch, loss_type)
  rsme = lstm_plot(yhat, history, test_X, test_y, airline, loss_type)
  return rsme

lstm_rsme_dict50 = {}

for airline in lstm_g.groups.keys():
  print(airline)
  n_epochs = 50
  n_batch = 72
  loss_type = 'mean_squared_error'
  model_rsme = lstm_run(airline, n_epochs, n_batch, loss_type)
  feat_lst = [model_rsme, n_epochs, n_batch]
  lstm_rsme_dict50[airline] = feat_lst

AA
Epoch 1/50
2/2 - 3s - loss: 0.2751 - val_loss: 0.3122 - 3s/epoch - 1s/step
Epoch 2/50
2/2 - 0s - loss: 0.1069 - val_loss: 0.1265 - 41ms/epoch - 20ms/step
Epoch 3/50
2/2 - 0s - loss: 0.0763 - val_loss: 0.0549 - 41ms/epoch - 20ms/step
Epoch 4/50
2/2 - 0s - loss: 0.1156 - val_loss: 0.0407 - 42ms/epoch - 21ms/step
Epoch 5/50
2/2 - 0s - loss: 0.1410 - val_loss: 0.0427 - 44ms/epoch - 22ms/step
Epoch 6/50
2/2 - 0s - loss: 0.1293 - val_loss: 0.0548 - 39ms/epoch - 19ms/step
Epoch 7/50
2/2 - 0s - loss: 0.1008 - val_loss: 0.0793 - 40ms/epoch - 20ms/step
Epoch 8/50
2/2 - 0s - loss: 0.0766 - val_loss: 0.1125 - 38ms/epoch - 19ms/step
Epoch 9/50
2/2 - 0s - loss: 0.0645 - val_loss: 0.1450 - 40ms/epoch - 20ms/step
Epoch 10/50
2/2 - 0s - loss: 0.0618 - val_loss: 0.1674 - 39ms/epoch - 19ms/step
Epoch 11/50
2/2 - 0s - loss: 0.0620 - val_loss: 0.1748 - 43ms/epoch - 22ms/step
Epoch 12/50
2/2 - 0s - loss: 0.0604 - val_loss: 0.1682 - 43ms/epoch - 22ms/step
Epoch 13/50
2/2 - 0s - loss: 0.0563 - val_loss: 0.1524 - 44ms/epoch - 22ms/step
Epoch 14/50
2/2 - 0s - loss: 0.0513 - val_loss: 0.1339 - 40ms/epoch - 20ms/step
Epoch 15/50
2/2 - 0s - loss: 0.0475 - val_loss: 0.1179 - 42ms/epoch - 21ms/step
Epoch 16/50
2/2 - 0s - loss: 0.0459 - val_loss: 0.1076 - 42ms/epoch - 21ms/step
Epoch 17/50
2/2 - 0s - loss: 0.0454 - val_loss: 0.1040 - 44ms/epoch - 22ms/step
Epoch 18/50
2/2 - 0s - loss: 0.0442 - val_loss: 0.1068 - 42ms/epoch - 21ms/step
Epoch 19/50
2/2 - 0s - loss: 0.0417 - val_loss: 0.1150 - 52ms/epoch - 26ms/step
Epoch 20/50
2/2 - 0s - loss: 0.0383 - val_loss: 0.1268 - 38ms/epoch - 19ms/step
Epoch 21/50
2/2 - 0s - loss: 0.0355 - val_loss: 0.1396 - 41ms/epoch - 21ms/step
Epoch 22/50
2/2 - 0s - loss: 0.0338 - val_loss: 0.1502 - 40ms/epoch - 20ms/step
Epoch 23/50
2/2 - 0s - loss: 0.0326 - val_loss: 0.1559 - 40ms/epoch - 20ms/step
Epoch 24/50
2/2 - 0s - loss: 0.0312 - val_loss: 0.1562 - 39ms/epoch - 20ms/step
Epoch 25/50
2/2 - 0s - loss: 0.0291 - val_loss: 0.1521 - 40ms/epoch - 20ms/step
Epoch 26/50
2/2 - 0s - loss: 0.0268 - val_loss: 0.1463 - 41ms/epoch - 20ms/step
Epoch 27/50
2/2 - 0s - loss: 0.0246 - val_loss: 0.1415 - 42ms/epoch - 21ms/step
Epoch 28/50
2/2 - 0s - loss: 0.0228 - val_loss: 0.1396 - 40ms/epoch - 20ms/step
Epoch 29/50
2/2 - 0s - loss: 0.0213 - val_loss: 0.1414 - 39ms/epoch - 19ms/step
Epoch 30/50
2/2 - 0s - loss: 0.0196 - val_loss: 0.1465 - 40ms/epoch - 20ms/step
Epoch 31/50
2/2 - 0s - loss: 0.0180 - val_loss: 0.1534 - 38ms/epoch - 19ms/step
Epoch 32/50
2/2 - 0s - loss: 0.0166 - val_loss: 0.1601 - 42ms/epoch - 21ms/step
Epoch 33/50
2/2 - 0s - loss: 0.0154 - val_loss: 0.1645 - 39ms/epoch - 20ms/step
Epoch 34/50
2/2 - 0s - loss: 0.0142 - val_loss: 0.1660 - 43ms/epoch - 22ms/step
Epoch 35/50
2/2 - 0s - loss: 0.0130 - val_loss: 0.1650 - 39ms/epoch - 20ms/step
Epoch 36/50
2/2 - 0s - loss: 0.0118 - val_loss: 0.1635 - 39ms/epoch - 20ms/step
Epoch 37/50
2/2 - 0s - loss: 0.0108 - val_loss: 0.1631 - 42ms/epoch - 21ms/step
Epoch 38/50
2/2 - 0s - loss: 0.0099 - val_loss: 0.1651 - 39ms/epoch - 19ms/step
Epoch 39/50
2/2 - 0s - loss: 0.0090 - val_loss: 0.1693 - 40ms/epoch - 20ms/step
Epoch 40/50
2/2 - 0s - loss: 0.0082 - val_loss: 0.1745 - 40ms/epoch - 20ms/step
Epoch 41/50
2/2 - 0s - loss: 0.0075 - val_loss: 0.1791 - 39ms/epoch - 20ms/step
Epoch 42/50
2/2 - 0s - loss: 0.0070 - val_loss: 0.1817 - 42ms/epoch - 21ms/step
Epoch 43/50
2/2 - 0s - loss: 0.0064 - val_loss: 0.1825 - 40ms/epoch - 20ms/step
Epoch 44/50
2/2 - 0s - loss: 0.0059 - val_loss: 0.1825 - 41ms/epoch - 20ms/step
Epoch 45/50
2/2 - 0s - loss: 0.0055 - val_loss: 0.1832 - 41ms/epoch - 20ms/step
Epoch 46/50
2/2 - 0s - loss: 0.0051 - val_loss: 0.1853 - 41ms/epoch - 20ms/step
Epoch 47/50
2/2 - 0s - loss: 0.0048 - val_loss: 0.1885 - 39ms/epoch - 20ms/step
Epoch 48/50
2/2 - 0s - loss: 0.0045 - val_loss: 0.1917 - 41ms/epoch - 21ms/step
Epoch 49/50
2/2 - 0s - loss: 0.0042 - val_loss: 0.1937 - 41ms/epoch - 20ms/step
Epoch 50/50
2/2 - 0s - loss: 0.0040 - val_loss: 0.1943 - 39ms/epoch - 19ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.441
AS
Epoch 1/50
2/2 - 3s - loss: 0.2292 - val_loss: 0.2681 - 3s/epoch - 1s/step
Epoch 2/50
2/2 - 0s - loss: 0.1045 - val_loss: 0.1762 - 38ms/epoch - 19ms/step
Epoch 3/50
2/2 - 0s - loss: 0.1041 - val_loss: 0.1437 - 39ms/epoch - 20ms/step
Epoch 4/50
2/2 - 0s - loss: 0.1507 - val_loss: 0.1359 - 40ms/epoch - 20ms/step
Epoch 5/50
2/2 - 0s - loss: 0.1747 - val_loss: 0.1336 - 38ms/epoch - 19ms/step
Epoch 6/50
2/2 - 0s - loss: 0.1632 - val_loss: 0.1348 - 38ms/epoch - 19ms/step
Epoch 7/50
2/2 - 0s - loss: 0.1343 - val_loss: 0.1413 - 40ms/epoch - 20ms/step
Epoch 8/50
2/2 - 0s - loss: 0.1060 - val_loss: 0.1521 - 40ms/epoch - 20ms/step
Epoch 9/50
2/2 - 0s - loss: 0.0866 - val_loss: 0.1633 - 40ms/epoch - 20ms/step
Epoch 10/50
2/2 - 0s - loss: 0.0760 - val_loss: 0.1709 - 39ms/epoch - 20ms/step
Epoch 11/50
2/2 - 0s - loss: 0.0703 - val_loss: 0.1723 - 39ms/epoch - 20ms/step
Epoch 12/50
2/2 - 0s - loss: 0.0661 - val_loss: 0.1675 - 36ms/epoch - 18ms/step
Epoch 13/50
2/2 - 0s - loss: 0.0617 - val_loss: 0.1586 - 40ms/epoch - 20ms/step
Epoch 14/50
2/2 - 0s - loss: 0.0576 - val_loss: 0.1483 - 42ms/epoch - 21ms/step
Epoch 15/50
2/2 - 0s - loss: 0.0548 - val_loss: 0.1392 - 41ms/epoch - 20ms/step
Epoch 16/50
2/2 - 0s - loss: 0.0535 - val_loss: 0.1327 - 39ms/epoch - 19ms/step
Epoch 17/50
2/2 - 0s - loss: 0.0528 - val_loss: 0.1289 - 38ms/epoch - 19ms/step
Epoch 18/50
2/2 - 0s - loss: 0.0509 - val_loss: 0.1278 - 40ms/epoch - 20ms/step
Epoch 19/50
2/2 - 0s - loss: 0.0470 - val_loss: 0.1288 - 39ms/epoch - 20ms/step
Epoch 20/50
2/2 - 0s - loss: 0.0417 - val_loss: 0.1315 - 38ms/epoch - 19ms/step
Epoch 21/50
2/2 - 0s - loss: 0.0364 - val_loss: 0.1351 - 36ms/epoch - 18ms/step
Epoch 22/50
2/2 - 0s - loss: 0.0324 - val_loss: 0.1382 - 39ms/epoch - 19ms/step
Epoch 23/50
2/2 - 0s - loss: 0.0298 - val_loss: 0.1393 - 40ms/epoch - 20ms/step
Epoch 24/50
2/2 - 0s - loss: 0.0277 - val_loss: 0.1381 - 39ms/epoch - 19ms/step
Epoch 25/50
2/2 - 0s - loss: 0.0255 - val_loss: 0.1353 - 40ms/epoch - 20ms/step
Epoch 26/50
2/2 - 0s - loss: 0.0234 - val_loss: 0.1323 - 40ms/epoch - 20ms/step
Epoch 27/50
2/2 - 0s - loss: 0.0216 - val_loss: 0.1305 - 39ms/epoch - 20ms/step
Epoch 28/50
2/2 - 0s - loss: 0.0201 - val_loss: 0.1305 - 41ms/epoch - 20ms/step
Epoch 29/50
2/2 - 0s - loss: 0.0187 - val_loss: 0.1322 - 39ms/epoch - 19ms/step
Epoch 30/50
2/2 - 0s - loss: 0.0173 - val_loss: 0.1351 - 39ms/epoch - 20ms/step
Epoch 31/50
2/2 - 0s - loss: 0.0163 - val_loss: 0.1379 - 39ms/epoch - 19ms/step
Epoch 32/50
2/2 - 0s - loss: 0.0157 - val_loss: 0.1393 - 39ms/epoch - 19ms/step
Epoch 33/50
2/2 - 0s - loss: 0.0150 - val_loss: 0.1391 - 43ms/epoch - 21ms/step
Epoch 34/50
2/2 - 0s - loss: 0.0139 - val_loss: 0.1381 - 36ms/epoch - 18ms/step
Epoch 35/50
2/2 - 0s - loss: 0.0127 - val_loss: 0.1373 - 40ms/epoch - 20ms/step
Epoch 36/50
2/2 - 0s - loss: 0.0117 - val_loss: 0.1377 - 38ms/epoch - 19ms/step
Epoch 37/50
2/2 - 0s - loss: 0.0109 - val_loss: 0.1393 - 40ms/epoch - 20ms/step
Epoch 38/50
2/2 - 0s - loss: 0.0102 - val_loss: 0.1417 - 38ms/epoch - 19ms/step
Epoch 39/50
2/2 - 0s - loss: 0.0096 - val_loss: 0.1436 - 40ms/epoch - 20ms/step
Epoch 40/50
2/2 - 0s - loss: 0.0091 - val_loss: 0.1445 - 41ms/epoch - 20ms/step
Epoch 41/50
2/2 - 0s - loss: 0.0085 - val_loss: 0.1445 - 39ms/epoch - 20ms/step
Epoch 42/50
2/2 - 0s - loss: 0.0078 - val_loss: 0.1442 - 41ms/epoch - 20ms/step
Epoch 43/50
2/2 - 0s - loss: 0.0072 - val_loss: 0.1446 - 43ms/epoch - 21ms/step
Epoch 44/50
2/2 - 0s - loss: 0.0067 - val_loss: 0.1458 - 41ms/epoch - 20ms/step
Epoch 45/50
2/2 - 0s - loss: 0.0063 - val_loss: 0.1476 - 42ms/epoch - 21ms/step
Epoch 46/50
2/2 - 0s - loss: 0.0059 - val_loss: 0.1492 - 40ms/epoch - 20ms/step
Epoch 47/50
2/2 - 0s - loss: 0.0056 - val_loss: 0.1501 - 40ms/epoch - 20ms/step
Epoch 48/50
2/2 - 0s - loss: 0.0053 - val_loss: 0.1504 - 40ms/epoch - 20ms/step
Epoch 49/50
2/2 - 0s - loss: 0.0050 - val_loss: 0.1506 - 42ms/epoch - 21ms/step
Epoch 50/50
2/2 - 0s - loss: 0.0047 - val_loss: 0.1513 - 41ms/epoch - 21ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.389
B6
Epoch 1/50
2/2 - 3s - loss: 0.3249 - val_loss: 0.1976 - 3s/epoch - 1s/step
Epoch 2/50
2/2 - 0s - loss: 0.1353 - val_loss: 0.0787 - 40ms/epoch - 20ms/step
Epoch 3/50
2/2 - 0s - loss: 0.0551 - val_loss: 0.0593 - 44ms/epoch - 22ms/step
Epoch 4/50
2/2 - 0s - loss: 0.0619 - val_loss: 0.0906 - 38ms/epoch - 19ms/step
Epoch 5/50
2/2 - 0s - loss: 0.1049 - val_loss: 0.1106 - 37ms/epoch - 18ms/step
Epoch 6/50
2/2 - 0s - loss: 0.1269 - val_loss: 0.1013 - 40ms/epoch - 20ms/step
Epoch 7/50
2/2 - 0s - loss: 0.1154 - val_loss: 0.0795 - 39ms/epoch - 20ms/step
Epoch 8/50
2/2 - 0s - loss: 0.0880 - val_loss: 0.0639 - 41ms/epoch - 21ms/step
Epoch 9/50
2/2 - 0s - loss: 0.0630 - val_loss: 0.0612 - 40ms/epoch - 20ms/step
Epoch 10/50
2/2 - 0s - loss: 0.0487 - val_loss: 0.0679 - 40ms/epoch - 20ms/step
Epoch 11/50
2/2 - 0s - loss: 0.0442 - val_loss: 0.0764 - 41ms/epoch - 20ms/step
Epoch 12/50
2/2 - 0s - loss: 0.0444 - val_loss: 0.0811 - 40ms/epoch - 20ms/step
Epoch 13/50
2/2 - 0s - loss: 0.0446 - val_loss: 0.0801 - 40ms/epoch - 20ms/step
Epoch 14/50
2/2 - 0s - loss: 0.0432 - val_loss: 0.0753 - 42ms/epoch - 21ms/step
Epoch 15/50
2/2 - 0s - loss: 0.0408 - val_loss: 0.0699 - 40ms/epoch - 20ms/step
Epoch 16/50
2/2 - 0s - loss: 0.0392 - val_loss: 0.0663 - 40ms/epoch - 20ms/step
Epoch 17/50
2/2 - 0s - loss: 0.0394 - val_loss: 0.0652 - 41ms/epoch - 20ms/step
Epoch 18/50
2/2 - 0s - loss: 0.0410 - val_loss: 0.0659 - 41ms/epoch - 20ms/step
Epoch 19/50
2/2 - 0s - loss: 0.0428 - val_loss: 0.0667 - 42ms/epoch - 21ms/step
Epoch 20/50
2/2 - 0s - loss: 0.0431 - val_loss: 0.0670 - 41ms/epoch - 21ms/step
Epoch 21/50
2/2 - 0s - loss: 0.0414 - val_loss: 0.0671 - 41ms/epoch - 21ms/step
Epoch 22/50
2/2 - 0s - loss: 0.0386 - val_loss: 0.0678 - 40ms/epoch - 20ms/step
Epoch 23/50
2/2 - 0s - loss: 0.0357 - val_loss: 0.0692 - 40ms/epoch - 20ms/step
Epoch 24/50
2/2 - 0s - loss: 0.0335 - val_loss: 0.0711 - 41ms/epoch - 21ms/step
Epoch 25/50
2/2 - 0s - loss: 0.0321 - val_loss: 0.0728 - 43ms/epoch - 22ms/step
Epoch 26/50
2/2 - 0s - loss: 0.0313 - val_loss: 0.0737 - 41ms/epoch - 21ms/step
Epoch 27/50
2/2 - 0s - loss: 0.0305 - val_loss: 0.0738 - 40ms/epoch - 20ms/step
Epoch 28/50
2/2 - 0s - loss: 0.0297 - val_loss: 0.0733 - 40ms/epoch - 20ms/step
Epoch 29/50
2/2 - 0s - loss: 0.0289 - val_loss: 0.0728 - 45ms/epoch - 23ms/step
Epoch 30/50
2/2 - 0s - loss: 0.0283 - val_loss: 0.0726 - 40ms/epoch - 20ms/step
Epoch 31/50
2/2 - 0s - loss: 0.0279 - val_loss: 0.0728 - 41ms/epoch - 20ms/step
Epoch 32/50
2/2 - 0s - loss: 0.0273 - val_loss: 0.0733 - 39ms/epoch - 20ms/step
Epoch 33/50
2/2 - 0s - loss: 0.0266 - val_loss: 0.0742 - 39ms/epoch - 19ms/step
Epoch 34/50
2/2 - 0s - loss: 0.0257 - val_loss: 0.0755 - 42ms/epoch - 21ms/step
Epoch 35/50
2/2 - 0s - loss: 0.0247 - val_loss: 0.0770 - 40ms/epoch - 20ms/step
Epoch 36/50
2/2 - 0s - loss: 0.0238 - val_loss: 0.0786 - 38ms/epoch - 19ms/step
Epoch 37/50
2/2 - 0s - loss: 0.0230 - val_loss: 0.0799 - 37ms/epoch - 18ms/step
Epoch 38/50
2/2 - 0s - loss: 0.0223 - val_loss: 0.0808 - 38ms/epoch - 19ms/step
Epoch 39/50
2/2 - 0s - loss: 0.0216 - val_loss: 0.0813 - 41ms/epoch - 20ms/step
Epoch 40/50
2/2 - 0s - loss: 0.0210 - val_loss: 0.0817 - 38ms/epoch - 19ms/step
Epoch 41/50
2/2 - 0s - loss: 0.0203 - val_loss: 0.0821 - 38ms/epoch - 19ms/step
Epoch 42/50
2/2 - 0s - loss: 0.0197 - val_loss: 0.0827 - 40ms/epoch - 20ms/step
Epoch 43/50
2/2 - 0s - loss: 0.0192 - val_loss: 0.0838 - 41ms/epoch - 20ms/step
Epoch 44/50
2/2 - 0s - loss: 0.0186 - val_loss: 0.0851 - 42ms/epoch - 21ms/step
Epoch 45/50
2/2 - 0s - loss: 0.0180 - val_loss: 0.0867 - 42ms/epoch - 21ms/step
Epoch 46/50
2/2 - 0s - loss: 0.0174 - val_loss: 0.0883 - 40ms/epoch - 20ms/step
Epoch 47/50
2/2 - 0s - loss: 0.0169 - val_loss: 0.0898 - 41ms/epoch - 21ms/step
Epoch 48/50
2/2 - 0s - loss: 0.0164 - val_loss: 0.0910 - 40ms/epoch - 20ms/step
Epoch 49/50
2/2 - 0s - loss: 0.0159 - val_loss: 0.0919 - 54ms/epoch - 27ms/step
Epoch 50/50
2/2 - 0s - loss: 0.0154 - val_loss: 0.0928 - 39ms/epoch - 19ms/step
4/4 [==============================] - 0s 4ms/step

Test RMSE: 0.305
DL
Epoch 1/50
1/1 - 2s - loss: 0.7080 - val_loss: 0.5901 - 2s/epoch - 2s/step
Epoch 2/50
1/1 - 0s - loss: 0.5182 - val_loss: 0.4108 - 37ms/epoch - 37ms/step
Epoch 3/50
1/1 - 0s - loss: 0.3629 - val_loss: 0.2714 - 36ms/epoch - 36ms/step
Epoch 4/50
1/1 - 0s - loss: 0.2427 - val_loss: 0.1711 - 36ms/epoch - 36ms/step
Epoch 5/50
1/1 - 0s - loss: 0.1571 - val_loss: 0.1077 - 37ms/epoch - 37ms/step
Epoch 6/50
1/1 - 0s - loss: 0.1044 - val_loss: 0.0770 - 38ms/epoch - 38ms/step
Epoch 7/50
1/1 - 0s - loss: 0.0812 - val_loss: 0.0725 - 37ms/epoch - 37ms/step
Epoch 8/50
1/1 - 0s - loss: 0.0817 - val_loss: 0.0845 - 36ms/epoch - 36ms/step
Epoch 9/50
1/1 - 0s - loss: 0.0971 - val_loss: 0.1023 - 37ms/epoch - 37ms/step
Epoch 10/50
1/1 - 0s - loss: 0.1175 - val_loss: 0.1172 - 37ms/epoch - 37ms/step
Epoch 11/50
1/1 - 0s - loss: 0.1346 - val_loss: 0.1246 - 37ms/epoch - 37ms/step
Epoch 12/50
1/1 - 0s - loss: 0.1439 - val_loss: 0.1235 - 38ms/epoch - 38ms/step
Epoch 13/50
1/1 - 0s - loss: 0.1445 - val_loss: 0.1156 - 37ms/epoch - 37ms/step
Epoch 14/50
1/1 - 0s - loss: 0.1376 - val_loss: 0.1034 - 37ms/epoch - 37ms/step
Epoch 15/50
1/1 - 0s - loss: 0.1257 - val_loss: 0.0895 - 35ms/epoch - 35ms/step
Epoch 16/50
1/1 - 0s - loss: 0.1114 - val_loss: 0.0764 - 36ms/epoch - 36ms/step
Epoch 17/50
1/1 - 0s - loss: 0.0969 - val_loss: 0.0656 - 37ms/epoch - 37ms/step
Epoch 18/50
1/1 - 0s - loss: 0.0839 - val_loss: 0.0582 - 36ms/epoch - 36ms/step
Epoch 19/50
1/1 - 0s - loss: 0.0736 - val_loss: 0.0545 - 38ms/epoch - 38ms/step
Epoch 20/50
1/1 - 0s - loss: 0.0665 - val_loss: 0.0543 - 37ms/epoch - 37ms/step
Epoch 21/50
1/1 - 0s - loss: 0.0624 - val_loss: 0.0570 - 37ms/epoch - 37ms/step
Epoch 22/50
1/1 - 0s - loss: 0.0610 - val_loss: 0.0615 - 36ms/epoch - 36ms/step
Epoch 23/50
1/1 - 0s - loss: 0.0614 - val_loss: 0.0670 - 39ms/epoch - 39ms/step
Epoch 24/50
1/1 - 0s - loss: 0.0631 - val_loss: 0.0725 - 34ms/epoch - 34ms/step
Epoch 25/50
1/1 - 0s - loss: 0.0652 - val_loss: 0.0774 - 37ms/epoch - 37ms/step
Epoch 26/50
1/1 - 0s - loss: 0.0671 - val_loss: 0.0809 - 37ms/epoch - 37ms/step
Epoch 27/50
1/1 - 0s - loss: 0.0684 - val_loss: 0.0830 - 36ms/epoch - 36ms/step
Epoch 28/50
1/1 - 0s - loss: 0.0688 - val_loss: 0.0833 - 36ms/epoch - 36ms/step
Epoch 29/50
1/1 - 0s - loss: 0.0683 - val_loss: 0.0821 - 38ms/epoch - 38ms/step
Epoch 30/50
1/1 - 0s - loss: 0.0669 - val_loss: 0.0795 - 38ms/epoch - 38ms/step
Epoch 31/50
1/1 - 0s - loss: 0.0647 - val_loss: 0.0759 - 37ms/epoch - 37ms/step
Epoch 32/50
1/1 - 0s - loss: 0.0622 - val_loss: 0.0716 - 39ms/epoch - 39ms/step
Epoch 33/50
1/1 - 0s - loss: 0.0594 - val_loss: 0.0669 - 37ms/epoch - 37ms/step
Epoch 34/50
1/1 - 0s - loss: 0.0567 - val_loss: 0.0623 - 35ms/epoch - 35ms/step
Epoch 35/50
1/1 - 0s - loss: 0.0543 - val_loss: 0.0579 - 36ms/epoch - 36ms/step
Epoch 36/50
1/1 - 0s - loss: 0.0524 - val_loss: 0.0539 - 35ms/epoch - 35ms/step
Epoch 37/50
1/1 - 0s - loss: 0.0510 - val_loss: 0.0505 - 36ms/epoch - 36ms/step
Epoch 38/50
1/1 - 0s - loss: 0.0501 - val_loss: 0.0478 - 36ms/epoch - 36ms/step
Epoch 39/50
1/1 - 0s - loss: 0.0497 - val_loss: 0.0456 - 36ms/epoch - 36ms/step
Epoch 40/50
1/1 - 0s - loss: 0.0496 - val_loss: 0.0439 - 34ms/epoch - 34ms/step
Epoch 41/50
1/1 - 0s - loss: 0.0497 - val_loss: 0.0427 - 38ms/epoch - 38ms/step
Epoch 42/50
1/1 - 0s - loss: 0.0498 - val_loss: 0.0419 - 36ms/epoch - 36ms/step
Epoch 43/50
1/1 - 0s - loss: 0.0498 - val_loss: 0.0414 - 37ms/epoch - 37ms/step
Epoch 44/50
1/1 - 0s - loss: 0.0496 - val_loss: 0.0412 - 39ms/epoch - 39ms/step
Epoch 45/50
1/1 - 0s - loss: 0.0492 - val_loss: 0.0413 - 38ms/epoch - 38ms/step
Epoch 46/50
1/1 - 0s - loss: 0.0485 - val_loss: 0.0417 - 37ms/epoch - 37ms/step
Epoch 47/50
1/1 - 0s - loss: 0.0477 - val_loss: 0.0423 - 38ms/epoch - 38ms/step
Epoch 48/50
1/1 - 0s - loss: 0.0468 - val_loss: 0.0433 - 37ms/epoch - 37ms/step
Epoch 49/50
1/1 - 0s - loss: 0.0458 - val_loss: 0.0446 - 35ms/epoch - 35ms/step
Epoch 50/50
1/1 - 0s - loss: 0.0449 - val_loss: 0.0462 - 37ms/epoch - 37ms/step
4/4 [==============================] - 0s 3ms/step

Test RMSE: 0.215
G4
Epoch 1/50
1/1 - 6s - loss: 0.0633 - val_loss: 0.3544 - 6s/epoch - 6s/step
Epoch 2/50
1/1 - 0s - loss: 0.0267 - val_loss: 0.2877 - 38ms/epoch - 38ms/step
Epoch 3/50
1/1 - 0s - loss: 0.0121 - val_loss: 0.2459 - 38ms/epoch - 38ms/step
Epoch 4/50
1/1 - 0s - loss: 0.0137 - val_loss: 0.2254 - 36ms/epoch - 36ms/step
Epoch 5/50
1/1 - 0s - loss: 0.0206 - val_loss: 0.2189 - 36ms/epoch - 36ms/step
Epoch 6/50
1/1 - 0s - loss: 0.0241 - val_loss: 0.2210 - 36ms/epoch - 36ms/step
Epoch 7/50
1/1 - 0s - loss: 0.0225 - val_loss: 0.2294 - 35ms/epoch - 35ms/step
Epoch 8/50
1/1 - 0s - loss: 0.0178 - val_loss: 0.2429 - 37ms/epoch - 37ms/step
Epoch 9/50
1/1 - 0s - loss: 0.0126 - val_loss: 0.2605 - 37ms/epoch - 37ms/step
Epoch 10/50
1/1 - 0s - loss: 0.0088 - val_loss: 0.2804 - 38ms/epoch - 38ms/step
Epoch 11/50
1/1 - 0s - loss: 0.0071 - val_loss: 0.3001 - 38ms/epoch - 38ms/step
Epoch 12/50
1/1 - 0s - loss: 0.0073 - val_loss: 0.3172 - 40ms/epoch - 40ms/step
Epoch 13/50
1/1 - 0s - loss: 0.0085 - val_loss: 0.3295 - 38ms/epoch - 38ms/step
Epoch 14/50
1/1 - 0s - loss: 0.0097 - val_loss: 0.3357 - 36ms/epoch - 36ms/step
Epoch 15/50
1/1 - 0s - loss: 0.0102 - val_loss: 0.3359 - 37ms/epoch - 37ms/step
Epoch 16/50
1/1 - 0s - loss: 0.0096 - val_loss: 0.3308 - 38ms/epoch - 38ms/step
Epoch 17/50
1/1 - 0s - loss: 0.0083 - val_loss: 0.3221 - 38ms/epoch - 38ms/step
Epoch 18/50
1/1 - 0s - loss: 0.0067 - val_loss: 0.3112 - 36ms/epoch - 36ms/step
Epoch 19/50
1/1 - 0s - loss: 0.0052 - val_loss: 0.2997 - 38ms/epoch - 38ms/step
Epoch 20/50
1/1 - 0s - loss: 0.0043 - val_loss: 0.2891 - 36ms/epoch - 36ms/step
Epoch 21/50
1/1 - 0s - loss: 0.0040 - val_loss: 0.2802 - 39ms/epoch - 39ms/step
Epoch 22/50
1/1 - 0s - loss: 0.0041 - val_loss: 0.2738 - 36ms/epoch - 36ms/step
Epoch 23/50
1/1 - 0s - loss: 0.0045 - val_loss: 0.2701 - 36ms/epoch - 36ms/step
Epoch 24/50
1/1 - 0s - loss: 0.0048 - val_loss: 0.2691 - 38ms/epoch - 38ms/step
Epoch 25/50
1/1 - 0s - loss: 0.0048 - val_loss: 0.2709 - 36ms/epoch - 36ms/step
Epoch 26/50
1/1 - 0s - loss: 0.0045 - val_loss: 0.2750 - 38ms/epoch - 38ms/step
Epoch 27/50
1/1 - 0s - loss: 0.0039 - val_loss: 0.2811 - 38ms/epoch - 38ms/step
Epoch 28/50
1/1 - 0s - loss: 0.0032 - val_loss: 0.2885 - 37ms/epoch - 37ms/step
Epoch 29/50
1/1 - 0s - loss: 0.0027 - val_loss: 0.2964 - 36ms/epoch - 36ms/step
Epoch 30/50
1/1 - 0s - loss: 0.0024 - val_loss: 0.3040 - 36ms/epoch - 36ms/step
Epoch 31/50
1/1 - 0s - loss: 0.0023 - val_loss: 0.3105 - 36ms/epoch - 36ms/step
Epoch 32/50
1/1 - 0s - loss: 0.0024 - val_loss: 0.3153 - 34ms/epoch - 34ms/step
Epoch 33/50
1/1 - 0s - loss: 0.0025 - val_loss: 0.3178 - 37ms/epoch - 37ms/step
Epoch 34/50
1/1 - 0s - loss: 0.0025 - val_loss: 0.3179 - 36ms/epoch - 36ms/step
Epoch 35/50
1/1 - 0s - loss: 0.0024 - val_loss: 0.3158 - 34ms/epoch - 34ms/step
Epoch 36/50
1/1 - 0s - loss: 0.0022 - val_loss: 0.3119 - 37ms/epoch - 37ms/step
Epoch 37/50
1/1 - 0s - loss: 0.0019 - val_loss: 0.3070 - 35ms/epoch - 35ms/step
Epoch 38/50
1/1 - 0s - loss: 0.0016 - val_loss: 0.3017 - 36ms/epoch - 36ms/step
Epoch 39/50
1/1 - 0s - loss: 0.0014 - val_loss: 0.2966 - 37ms/epoch - 37ms/step
Epoch 40/50
1/1 - 0s - loss: 0.0014 - val_loss: 0.2925 - 37ms/epoch - 37ms/step
Epoch 41/50
1/1 - 0s - loss: 0.0014 - val_loss: 0.2897 - 42ms/epoch - 42ms/step
Epoch 42/50
1/1 - 0s - loss: 0.0014 - val_loss: 0.2885 - 37ms/epoch - 37ms/step
Epoch 43/50
1/1 - 0s - loss: 0.0014 - val_loss: 0.2888 - 36ms/epoch - 36ms/step
Epoch 44/50
1/1 - 0s - loss: 0.0014 - val_loss: 0.2905 - 36ms/epoch - 36ms/step
Epoch 45/50
1/1 - 0s - loss: 0.0012 - val_loss: 0.2934 - 37ms/epoch - 37ms/step
Epoch 46/50
1/1 - 0s - loss: 0.0011 - val_loss: 0.2971 - 36ms/epoch - 36ms/step
Epoch 47/50
1/1 - 0s - loss: 9.9558e-04 - val_loss: 0.3009 - 34ms/epoch - 34ms/step
Epoch 48/50
1/1 - 0s - loss: 9.3669e-04 - val_loss: 0.3045 - 34ms/epoch - 34ms/step
Epoch 49/50
1/1 - 0s - loss: 9.2858e-04 - val_loss: 0.3072 - 37ms/epoch - 37ms/step
Epoch 50/50
1/1 - 0s - loss: 9.4507e-04 - val_loss: 0.3088 - 34ms/epoch - 34ms/step
2/2 [==============================] - 0s 5ms/step

Test RMSE: 0.556
HA
Epoch 1/50
1/1 - 3s - loss: 0.3960 - val_loss: 0.6918 - 3s/epoch - 3s/step
Epoch 2/50
1/1 - 0s - loss: 0.2906 - val_loss: 0.5265 - 36ms/epoch - 36ms/step
Epoch 3/50
1/1 - 0s - loss: 0.2107 - val_loss: 0.3928 - 34ms/epoch - 34ms/step
Epoch 4/50
1/1 - 0s - loss: 0.1557 - val_loss: 0.2902 - 37ms/epoch - 37ms/step
Epoch 5/50
1/1 - 0s - loss: 0.1241 - val_loss: 0.2173 - 33ms/epoch - 33ms/step
Epoch 6/50
1/1 - 0s - loss: 0.1126 - val_loss: 0.1706 - 35ms/epoch - 35ms/step
Epoch 7/50
1/1 - 0s - loss: 0.1155 - val_loss: 0.1443 - 36ms/epoch - 36ms/step
Epoch 8/50
1/1 - 0s - loss: 0.1252 - val_loss: 0.1310 - 34ms/epoch - 34ms/step
Epoch 9/50
1/1 - 0s - loss: 0.1345 - val_loss: 0.1251 - 37ms/epoch - 37ms/step
Epoch 10/50
1/1 - 0s - loss: 0.1388 - val_loss: 0.1235 - 35ms/epoch - 35ms/step
Epoch 11/50
1/1 - 0s - loss: 0.1368 - val_loss: 0.1252 - 34ms/epoch - 34ms/step
Epoch 12/50
1/1 - 0s - loss: 0.1295 - val_loss: 0.1305 - 35ms/epoch - 35ms/step
Epoch 13/50
1/1 - 0s - loss: 0.1187 - val_loss: 0.1401 - 38ms/epoch - 38ms/step
Epoch 14/50
1/1 - 0s - loss: 0.1068 - val_loss: 0.1543 - 34ms/epoch - 34ms/step
Epoch 15/50
1/1 - 0s - loss: 0.0955 - val_loss: 0.1728 - 35ms/epoch - 35ms/step
Epoch 16/50
1/1 - 0s - loss: 0.0860 - val_loss: 0.1951 - 36ms/epoch - 36ms/step
Epoch 17/50
1/1 - 0s - loss: 0.0791 - val_loss: 0.2197 - 36ms/epoch - 36ms/step
Epoch 18/50
1/1 - 0s - loss: 0.0746 - val_loss: 0.2451 - 35ms/epoch - 35ms/step
Epoch 19/50
1/1 - 0s - loss: 0.0722 - val_loss: 0.2698 - 34ms/epoch - 34ms/step
Epoch 20/50
1/1 - 0s - loss: 0.0712 - val_loss: 0.2922 - 38ms/epoch - 38ms/step
Epoch 21/50
1/1 - 0s - loss: 0.0707 - val_loss: 0.3110 - 35ms/epoch - 35ms/step
Epoch 22/50
1/1 - 0s - loss: 0.0702 - val_loss: 0.3252 - 35ms/epoch - 35ms/step
Epoch 23/50
1/1 - 0s - loss: 0.0690 - val_loss: 0.3344 - 35ms/epoch - 35ms/step
Epoch 24/50
1/1 - 0s - loss: 0.0668 - val_loss: 0.3384 - 35ms/epoch - 35ms/step
Epoch 25/50
1/1 - 0s - loss: 0.0636 - val_loss: 0.3376 - 35ms/epoch - 35ms/step
Epoch 26/50
1/1 - 0s - loss: 0.0596 - val_loss: 0.3326 - 37ms/epoch - 37ms/step
Epoch 27/50
1/1 - 0s - loss: 0.0551 - val_loss: 0.3244 - 36ms/epoch - 36ms/step
Epoch 28/50
1/1 - 0s - loss: 0.0505 - val_loss: 0.3141 - 36ms/epoch - 36ms/step
Epoch 29/50
1/1 - 0s - loss: 0.0463 - val_loss: 0.3030 - 35ms/epoch - 35ms/step
Epoch 30/50
1/1 - 0s - loss: 0.0427 - val_loss: 0.2923 - 37ms/epoch - 37ms/step
Epoch 31/50
1/1 - 0s - loss: 0.0400 - val_loss: 0.2831 - 36ms/epoch - 36ms/step
Epoch 32/50
1/1 - 0s - loss: 0.0381 - val_loss: 0.2765 - 37ms/epoch - 37ms/step
Epoch 33/50
1/1 - 0s - loss: 0.0367 - val_loss: 0.2733 - 38ms/epoch - 38ms/step
Epoch 34/50
1/1 - 0s - loss: 0.0357 - val_loss: 0.2742 - 37ms/epoch - 37ms/step
Epoch 35/50
1/1 - 0s - loss: 0.0346 - val_loss: 0.2796 - 40ms/epoch - 40ms/step
Epoch 36/50
1/1 - 0s - loss: 0.0331 - val_loss: 0.2897 - 36ms/epoch - 36ms/step
Epoch 37/50
1/1 - 0s - loss: 0.0313 - val_loss: 0.3043 - 37ms/epoch - 37ms/step
Epoch 38/50
1/1 - 0s - loss: 0.0291 - val_loss: 0.3233 - 37ms/epoch - 37ms/step
Epoch 39/50
1/1 - 0s - loss: 0.0268 - val_loss: 0.3458 - 36ms/epoch - 36ms/step
Epoch 40/50
1/1 - 0s - loss: 0.0246 - val_loss: 0.3707 - 34ms/epoch - 34ms/step
Epoch 41/50
1/1 - 0s - loss: 0.0228 - val_loss: 0.3968 - 33ms/epoch - 33ms/step
Epoch 42/50
1/1 - 0s - loss: 0.0215 - val_loss: 0.4223 - 36ms/epoch - 36ms/step
Epoch 43/50
1/1 - 0s - loss: 0.0207 - val_loss: 0.4455 - 36ms/epoch - 36ms/step
Epoch 44/50
1/1 - 0s - loss: 0.0201 - val_loss: 0.4651 - 35ms/epoch - 35ms/step
Epoch 45/50
1/1 - 0s - loss: 0.0196 - val_loss: 0.4799 - 36ms/epoch - 36ms/step
Epoch 46/50
1/1 - 0s - loss: 0.0190 - val_loss: 0.4894 - 39ms/epoch - 39ms/step
Epoch 47/50
1/1 - 0s - loss: 0.0181 - val_loss: 0.4939 - 39ms/epoch - 39ms/step
Epoch 48/50
1/1 - 0s - loss: 0.0171 - val_loss: 0.4941 - 36ms/epoch - 36ms/step
Epoch 49/50
1/1 - 0s - loss: 0.0161 - val_loss: 0.4913 - 39ms/epoch - 39ms/step
Epoch 50/50
1/1 - 0s - loss: 0.0151 - val_loss: 0.4872 - 36ms/epoch - 36ms/step
1/1 [==============================] - 0s 423ms/step

Test RMSE: 0.698
NK
Epoch 1/50
1/1 - 3s - loss: 0.5388 - val_loss: 0.3626 - 3s/epoch - 3s/step
Epoch 2/50
1/1 - 0s - loss: 0.4263 - val_loss: 0.2820 - 37ms/epoch - 37ms/step
Epoch 3/50
1/1 - 0s - loss: 0.3296 - val_loss: 0.2246 - 37ms/epoch - 37ms/step
Epoch 4/50
1/1 - 0s - loss: 0.2484 - val_loss: 0.1904 - 41ms/epoch - 41ms/step
Epoch 5/50
1/1 - 0s - loss: 0.1828 - val_loss: 0.1785 - 37ms/epoch - 37ms/step
Epoch 6/50
1/1 - 0s - loss: 0.1323 - val_loss: 0.1868 - 37ms/epoch - 37ms/step
Epoch 7/50
1/1 - 0s - loss: 0.0965 - val_loss: 0.2116 - 35ms/epoch - 35ms/step
Epoch 8/50
1/1 - 0s - loss: 0.0741 - val_loss: 0.2472 - 37ms/epoch - 37ms/step
Epoch 9/50
1/1 - 0s - loss: 0.0634 - val_loss: 0.2868 - 36ms/epoch - 36ms/step
Epoch 10/50
1/1 - 0s - loss: 0.0618 - val_loss: 0.3231 - 35ms/epoch - 35ms/step
Epoch 11/50
1/1 - 0s - loss: 0.0662 - val_loss: 0.3503 - 38ms/epoch - 38ms/step
Epoch 12/50
1/1 - 0s - loss: 0.0731 - val_loss: 0.3650 - 35ms/epoch - 35ms/step
Epoch 13/50
1/1 - 0s - loss: 0.0796 - val_loss: 0.3663 - 35ms/epoch - 35ms/step
Epoch 14/50
1/1 - 0s - loss: 0.0837 - val_loss: 0.3554 - 38ms/epoch - 38ms/step
Epoch 15/50
1/1 - 0s - loss: 0.0844 - val_loss: 0.3350 - 43ms/epoch - 43ms/step
Epoch 16/50
1/1 - 0s - loss: 0.0817 - val_loss: 0.3083 - 42ms/epoch - 42ms/step
Epoch 17/50
1/1 - 0s - loss: 0.0761 - val_loss: 0.2785 - 40ms/epoch - 40ms/step
Epoch 18/50
1/1 - 0s - loss: 0.0685 - val_loss: 0.2485 - 40ms/epoch - 40ms/step
Epoch 19/50
1/1 - 0s - loss: 0.0600 - val_loss: 0.2204 - 38ms/epoch - 38ms/step
Epoch 20/50
1/1 - 0s - loss: 0.0514 - val_loss: 0.1960 - 38ms/epoch - 38ms/step
Epoch 21/50
1/1 - 0s - loss: 0.0436 - val_loss: 0.1761 - 37ms/epoch - 37ms/step
Epoch 22/50
1/1 - 0s - loss: 0.0370 - val_loss: 0.1611 - 38ms/epoch - 38ms/step
Epoch 23/50
1/1 - 0s - loss: 0.0319 - val_loss: 0.1507 - 38ms/epoch - 38ms/step
Epoch 24/50
1/1 - 0s - loss: 0.0284 - val_loss: 0.1445 - 39ms/epoch - 39ms/step
Epoch 25/50
1/1 - 0s - loss: 0.0261 - val_loss: 0.1418 - 41ms/epoch - 41ms/step
Epoch 26/50
1/1 - 0s - loss: 0.0250 - val_loss: 0.1417 - 42ms/epoch - 42ms/step
Epoch 27/50
1/1 - 0s - loss: 0.0247 - val_loss: 0.1435 - 40ms/epoch - 40ms/step
Epoch 28/50
1/1 - 0s - loss: 0.0248 - val_loss: 0.1463 - 40ms/epoch - 40ms/step
Epoch 29/50
1/1 - 0s - loss: 0.0250 - val_loss: 0.1495 - 39ms/epoch - 39ms/step
Epoch 30/50
1/1 - 0s - loss: 0.0251 - val_loss: 0.1527 - 40ms/epoch - 40ms/step
Epoch 31/50
1/1 - 0s - loss: 0.0250 - val_loss: 0.1555 - 38ms/epoch - 38ms/step
Epoch 32/50
1/1 - 0s - loss: 0.0245 - val_loss: 0.1577 - 41ms/epoch - 41ms/step
Epoch 33/50
1/1 - 0s - loss: 0.0236 - val_loss: 0.1590 - 37ms/epoch - 37ms/step
Epoch 34/50
1/1 - 0s - loss: 0.0224 - val_loss: 0.1596 - 38ms/epoch - 38ms/step
Epoch 35/50
1/1 - 0s - loss: 0.0209 - val_loss: 0.1595 - 39ms/epoch - 39ms/step
Epoch 36/50
1/1 - 0s - loss: 0.0192 - val_loss: 0.1587 - 36ms/epoch - 36ms/step
Epoch 37/50
1/1 - 0s - loss: 0.0175 - val_loss: 0.1574 - 37ms/epoch - 37ms/step
Epoch 38/50
1/1 - 0s - loss: 0.0159 - val_loss: 0.1559 - 37ms/epoch - 37ms/step
Epoch 39/50
1/1 - 0s - loss: 0.0145 - val_loss: 0.1541 - 36ms/epoch - 36ms/step
Epoch 40/50
1/1 - 0s - loss: 0.0134 - val_loss: 0.1524 - 37ms/epoch - 37ms/step
Epoch 41/50
1/1 - 0s - loss: 0.0126 - val_loss: 0.1508 - 38ms/epoch - 38ms/step
Epoch 42/50
1/1 - 0s - loss: 0.0120 - val_loss: 0.1495 - 37ms/epoch - 37ms/step
Epoch 43/50
1/1 - 0s - loss: 0.0117 - val_loss: 0.1486 - 37ms/epoch - 37ms/step
Epoch 44/50
1/1 - 0s - loss: 0.0116 - val_loss: 0.1480 - 39ms/epoch - 39ms/step
Epoch 45/50
1/1 - 0s - loss: 0.0116 - val_loss: 0.1480 - 38ms/epoch - 38ms/step
Epoch 46/50
1/1 - 0s - loss: 0.0116 - val_loss: 0.1484 - 39ms/epoch - 39ms/step
Epoch 47/50
1/1 - 0s - loss: 0.0116 - val_loss: 0.1495 - 40ms/epoch - 40ms/step
Epoch 48/50
1/1 - 0s - loss: 0.0114 - val_loss: 0.1511 - 37ms/epoch - 37ms/step
Epoch 49/50
1/1 - 0s - loss: 0.0112 - val_loss: 0.1533 - 36ms/epoch - 36ms/step
Epoch 50/50
1/1 - 0s - loss: 0.0109 - val_loss: 0.1561 - 35ms/epoch - 35ms/step
3/3 [==============================] - 0s 4ms/step

Test RMSE: 0.395
UA
Epoch 1/50
2/2 - 3s - loss: 0.3827 - val_loss: 0.4592 - 3s/epoch - 1s/step
Epoch 2/50
2/2 - 0s - loss: 0.2386 - val_loss: 0.2808 - 41ms/epoch - 21ms/step
Epoch 3/50
2/2 - 0s - loss: 0.1957 - val_loss: 0.1700 - 39ms/epoch - 20ms/step
Epoch 4/50
2/2 - 0s - loss: 0.2138 - val_loss: 0.1119 - 42ms/epoch - 21ms/step
Epoch 5/50
2/2 - 0s - loss: 0.2531 - val_loss: 0.0866 - 40ms/epoch - 20ms/step
Epoch 6/50
2/2 - 0s - loss: 0.2792 - val_loss: 0.0776 - 42ms/epoch - 21ms/step
Epoch 7/50
2/2 - 0s - loss: 0.2791 - val_loss: 0.0770 - 41ms/epoch - 21ms/step
Epoch 8/50
2/2 - 0s - loss: 0.2582 - val_loss: 0.0819 - 41ms/epoch - 20ms/step
Epoch 9/50
2/2 - 0s - loss: 0.2283 - val_loss: 0.0911 - 44ms/epoch - 22ms/step
Epoch 10/50
2/2 - 0s - loss: 0.1990 - val_loss: 0.1024 - 44ms/epoch - 22ms/step
Epoch 11/50
2/2 - 0s - loss: 0.1749 - val_loss: 0.1132 - 42ms/epoch - 21ms/step
Epoch 12/50
2/2 - 0s - loss: 0.1569 - val_loss: 0.1209 - 43ms/epoch - 22ms/step
Epoch 13/50
2/2 - 0s - loss: 0.1439 - val_loss: 0.1237 - 41ms/epoch - 21ms/step
Epoch 14/50
2/2 - 0s - loss: 0.1340 - val_loss: 0.1211 - 43ms/epoch - 22ms/step
Epoch 15/50
2/2 - 0s - loss: 0.1260 - val_loss: 0.1138 - 42ms/epoch - 21ms/step
Epoch 16/50
2/2 - 0s - loss: 0.1192 - val_loss: 0.1035 - 44ms/epoch - 22ms/step
Epoch 17/50
2/2 - 0s - loss: 0.1138 - val_loss: 0.0920 - 39ms/epoch - 20ms/step
Epoch 18/50
2/2 - 0s - loss: 0.1096 - val_loss: 0.0812 - 49ms/epoch - 24ms/step
Epoch 19/50
2/2 - 0s - loss: 0.1067 - val_loss: 0.0724 - 40ms/epoch - 20ms/step
Epoch 20/50
2/2 - 0s - loss: 0.1042 - val_loss: 0.0661 - 43ms/epoch - 21ms/step
Epoch 21/50
2/2 - 0s - loss: 0.1011 - val_loss: 0.0622 - 43ms/epoch - 22ms/step
Epoch 22/50
2/2 - 0s - loss: 0.0964 - val_loss: 0.0603 - 43ms/epoch - 22ms/step
Epoch 23/50
2/2 - 0s - loss: 0.0898 - val_loss: 0.0600 - 56ms/epoch - 28ms/step
Epoch 24/50
2/2 - 0s - loss: 0.0817 - val_loss: 0.0606 - 41ms/epoch - 20ms/step
Epoch 25/50
2/2 - 0s - loss: 0.0732 - val_loss: 0.0617 - 41ms/epoch - 21ms/step
Epoch 26/50
2/2 - 0s - loss: 0.0652 - val_loss: 0.0624 - 45ms/epoch - 23ms/step
Epoch 27/50
2/2 - 0s - loss: 0.0584 - val_loss: 0.0623 - 42ms/epoch - 21ms/step
Epoch 28/50
2/2 - 0s - loss: 0.0529 - val_loss: 0.0611 - 42ms/epoch - 21ms/step
Epoch 29/50
2/2 - 0s - loss: 0.0484 - val_loss: 0.0593 - 55ms/epoch - 27ms/step
Epoch 30/50
2/2 - 0s - loss: 0.0448 - val_loss: 0.0574 - 42ms/epoch - 21ms/step
Epoch 31/50
2/2 - 0s - loss: 0.0421 - val_loss: 0.0560 - 42ms/epoch - 21ms/step
Epoch 32/50
2/2 - 0s - loss: 0.0399 - val_loss: 0.0554 - 40ms/epoch - 20ms/step
Epoch 33/50
2/2 - 0s - loss: 0.0376 - val_loss: 0.0554 - 41ms/epoch - 20ms/step
Epoch 34/50
2/2 - 0s - loss: 0.0351 - val_loss: 0.0559 - 41ms/epoch - 20ms/step
Epoch 35/50
2/2 - 0s - loss: 0.0324 - val_loss: 0.0566 - 42ms/epoch - 21ms/step
Epoch 36/50
2/2 - 0s - loss: 0.0298 - val_loss: 0.0573 - 42ms/epoch - 21ms/step
Epoch 37/50
2/2 - 0s - loss: 0.0276 - val_loss: 0.0577 - 41ms/epoch - 20ms/step
Epoch 38/50
2/2 - 0s - loss: 0.0259 - val_loss: 0.0577 - 39ms/epoch - 19ms/step
Epoch 39/50
2/2 - 0s - loss: 0.0247 - val_loss: 0.0574 - 48ms/epoch - 24ms/step
Epoch 40/50
2/2 - 0s - loss: 0.0238 - val_loss: 0.0570 - 43ms/epoch - 21ms/step
Epoch 41/50
2/2 - 0s - loss: 0.0232 - val_loss: 0.0568 - 41ms/epoch - 20ms/step
Epoch 42/50
2/2 - 0s - loss: 0.0225 - val_loss: 0.0568 - 40ms/epoch - 20ms/step
Epoch 43/50
2/2 - 0s - loss: 0.0217 - val_loss: 0.0570 - 41ms/epoch - 21ms/step
Epoch 44/50
2/2 - 0s - loss: 0.0209 - val_loss: 0.0571 - 49ms/epoch - 24ms/step
Epoch 45/50
2/2 - 0s - loss: 0.0200 - val_loss: 0.0572 - 41ms/epoch - 20ms/step
Epoch 46/50
2/2 - 0s - loss: 0.0193 - val_loss: 0.0570 - 42ms/epoch - 21ms/step
Epoch 47/50
2/2 - 0s - loss: 0.0188 - val_loss: 0.0567 - 43ms/epoch - 22ms/step
Epoch 48/50
2/2 - 0s - loss: 0.0184 - val_loss: 0.0564 - 43ms/epoch - 21ms/step
Epoch 49/50
2/2 - 0s - loss: 0.0180 - val_loss: 0.0561 - 42ms/epoch - 21ms/step
Epoch 50/50
2/2 - 0s - loss: 0.0176 - val_loss: 0.0560 - 43ms/epoch - 22ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.237
WN
Epoch 1/50
2/2 - 2s - loss: 0.5623 - val_loss: 0.1521 - 2s/epoch - 1s/step
Epoch 2/50
2/2 - 0s - loss: 0.3468 - val_loss: 0.1256 - 43ms/epoch - 22ms/step
Epoch 3/50
2/2 - 0s - loss: 0.2179 - val_loss: 0.1608 - 41ms/epoch - 21ms/step
Epoch 4/50
2/2 - 0s - loss: 0.1609 - val_loss: 0.2223 - 43ms/epoch - 21ms/step
Epoch 5/50
2/2 - 0s - loss: 0.1494 - val_loss: 0.2649 - 45ms/epoch - 22ms/step
Epoch 6/50
2/2 - 0s - loss: 0.1509 - val_loss: 0.2706 - 42ms/epoch - 21ms/step
Epoch 7/50
2/2 - 0s - loss: 0.1479 - val_loss: 0.2484 - 41ms/epoch - 20ms/step
Epoch 8/50
2/2 - 0s - loss: 0.1389 - val_loss: 0.2140 - 42ms/epoch - 21ms/step
Epoch 9/50
2/2 - 0s - loss: 0.1291 - val_loss: 0.1796 - 41ms/epoch - 21ms/step
Epoch 10/50
2/2 - 0s - loss: 0.1227 - val_loss: 0.1515 - 43ms/epoch - 22ms/step
Epoch 11/50
2/2 - 0s - loss: 0.1208 - val_loss: 0.1314 - 43ms/epoch - 21ms/step
Epoch 12/50
2/2 - 0s - loss: 0.1220 - val_loss: 0.1186 - 43ms/epoch - 22ms/step
Epoch 13/50
2/2 - 0s - loss: 0.1240 - val_loss: 0.1110 - 43ms/epoch - 22ms/step
Epoch 14/50
2/2 - 0s - loss: 0.1247 - val_loss: 0.1069 - 43ms/epoch - 21ms/step
Epoch 15/50
2/2 - 0s - loss: 0.1228 - val_loss: 0.1055 - 45ms/epoch - 23ms/step
Epoch 16/50
2/2 - 0s - loss: 0.1186 - val_loss: 0.1063 - 43ms/epoch - 21ms/step
Epoch 17/50
2/2 - 0s - loss: 0.1128 - val_loss: 0.1091 - 43ms/epoch - 22ms/step
Epoch 18/50
2/2 - 0s - loss: 0.1066 - val_loss: 0.1134 - 43ms/epoch - 22ms/step
Epoch 19/50
2/2 - 0s - loss: 0.1012 - val_loss: 0.1185 - 40ms/epoch - 20ms/step
Epoch 20/50
2/2 - 0s - loss: 0.0969 - val_loss: 0.1230 - 42ms/epoch - 21ms/step
Epoch 21/50
2/2 - 0s - loss: 0.0935 - val_loss: 0.1258 - 39ms/epoch - 19ms/step
Epoch 22/50
2/2 - 0s - loss: 0.0906 - val_loss: 0.1260 - 41ms/epoch - 21ms/step
Epoch 23/50
2/2 - 0s - loss: 0.0877 - val_loss: 0.1235 - 40ms/epoch - 20ms/step
Epoch 24/50
2/2 - 0s - loss: 0.0846 - val_loss: 0.1190 - 40ms/epoch - 20ms/step
Epoch 25/50
2/2 - 0s - loss: 0.0814 - val_loss: 0.1135 - 44ms/epoch - 22ms/step
Epoch 26/50
2/2 - 0s - loss: 0.0784 - val_loss: 0.1081 - 44ms/epoch - 22ms/step
Epoch 27/50
2/2 - 0s - loss: 0.0757 - val_loss: 0.1035 - 40ms/epoch - 20ms/step
Epoch 28/50
2/2 - 0s - loss: 0.0731 - val_loss: 0.1003 - 39ms/epoch - 20ms/step
Epoch 29/50
2/2 - 0s - loss: 0.0705 - val_loss: 0.0986 - 40ms/epoch - 20ms/step
Epoch 30/50
2/2 - 0s - loss: 0.0676 - val_loss: 0.0982 - 43ms/epoch - 22ms/step
Epoch 31/50
2/2 - 0s - loss: 0.0645 - val_loss: 0.0989 - 42ms/epoch - 21ms/step
Epoch 32/50
2/2 - 0s - loss: 0.0612 - val_loss: 0.1003 - 41ms/epoch - 20ms/step
Epoch 33/50
2/2 - 0s - loss: 0.0580 - val_loss: 0.1017 - 41ms/epoch - 21ms/step
Epoch 34/50
2/2 - 0s - loss: 0.0550 - val_loss: 0.1024 - 44ms/epoch - 22ms/step
Epoch 35/50
2/2 - 0s - loss: 0.0521 - val_loss: 0.1020 - 43ms/epoch - 21ms/step
Epoch 36/50
2/2 - 0s - loss: 0.0493 - val_loss: 0.1004 - 42ms/epoch - 21ms/step
Epoch 37/50
2/2 - 0s - loss: 0.0465 - val_loss: 0.0980 - 44ms/epoch - 22ms/step
Epoch 38/50
2/2 - 0s - loss: 0.0438 - val_loss: 0.0955 - 42ms/epoch - 21ms/step
Epoch 39/50
2/2 - 0s - loss: 0.0412 - val_loss: 0.0934 - 43ms/epoch - 22ms/step
Epoch 40/50
2/2 - 0s - loss: 0.0387 - val_loss: 0.0922 - 41ms/epoch - 21ms/step
Epoch 41/50
2/2 - 0s - loss: 0.0363 - val_loss: 0.0920 - 41ms/epoch - 21ms/step
Epoch 42/50
2/2 - 0s - loss: 0.0338 - val_loss: 0.0925 - 42ms/epoch - 21ms/step
Epoch 43/50
2/2 - 0s - loss: 0.0314 - val_loss: 0.0932 - 40ms/epoch - 20ms/step
Epoch 44/50
2/2 - 0s - loss: 0.0292 - val_loss: 0.0936 - 40ms/epoch - 20ms/step
Epoch 45/50
2/2 - 0s - loss: 0.0271 - val_loss: 0.0933 - 39ms/epoch - 20ms/step
Epoch 46/50
2/2 - 0s - loss: 0.0251 - val_loss: 0.0924 - 39ms/epoch - 19ms/step
Epoch 47/50
2/2 - 0s - loss: 0.0232 - val_loss: 0.0913 - 40ms/epoch - 20ms/step
Epoch 48/50
2/2 - 0s - loss: 0.0214 - val_loss: 0.0905 - 44ms/epoch - 22ms/step
Epoch 49/50
2/2 - 0s - loss: 0.0197 - val_loss: 0.0905 - 43ms/epoch - 21ms/step
Epoch 50/50
2/2 - 0s - loss: 0.0182 - val_loss: 0.0912 - 39ms/epoch - 20ms/step
5/5 [==============================] - 0s 4ms/step

Test RMSE: 0.302

lstm_rsme_dict20 = {}

for airline in lstm_g.groups.keys():
  print(airline)
  n_epochs = 20
  n_batch = 72
  loss_type = 'mean_squared_error'
  model_rsme = lstm_run(airline, n_epochs, n_batch, loss_type)
  feat_lst = [model_rsme, n_epochs, n_batch]
  lstm_rsme_dict20[airline] = feat_lst

AA
Epoch 1/20
2/2 - 3s - loss: 0.2505 - val_loss: 0.3238 - 3s/epoch - 1s/step
Epoch 2/20
2/2 - 0s - loss: 0.1569 - val_loss: 0.2063 - 41ms/epoch - 20ms/step
Epoch 3/20
2/2 - 0s - loss: 0.1033 - val_loss: 0.1209 - 39ms/epoch - 20ms/step
Epoch 4/20
2/2 - 0s - loss: 0.0830 - val_loss: 0.0675 - 39ms/epoch - 19ms/step
Epoch 5/20
2/2 - 0s - loss: 0.0881 - val_loss: 0.0414 - 39ms/epoch - 20ms/step
Epoch 6/20
2/2 - 0s - loss: 0.1037 - val_loss: 0.0323 - 39ms/epoch - 20ms/step
Epoch 7/20
2/2 - 0s - loss: 0.1142 - val_loss: 0.0313 - 40ms/epoch - 20ms/step
Epoch 8/20
2/2 - 0s - loss: 0.1135 - val_loss: 0.0346 - 39ms/epoch - 20ms/step
Epoch 9/20
2/2 - 0s - loss: 0.1042 - val_loss: 0.0418 - 38ms/epoch - 19ms/step
Epoch 10/20
2/2 - 0s - loss: 0.0919 - val_loss: 0.0526 - 40ms/epoch - 20ms/step
Epoch 11/20
2/2 - 0s - loss: 0.0808 - val_loss: 0.0658 - 41ms/epoch - 20ms/step
Epoch 12/20
2/2 - 0s - loss: 0.0730 - val_loss: 0.0792 - 39ms/epoch - 19ms/step
Epoch 13/20
2/2 - 0s - loss: 0.0682 - val_loss: 0.0905 - 38ms/epoch - 19ms/step
Epoch 14/20
2/2 - 0s - loss: 0.0656 - val_loss: 0.0982 - 38ms/epoch - 19ms/step
Epoch 15/20
2/2 - 0s - loss: 0.0638 - val_loss: 0.1016 - 39ms/epoch - 19ms/step
Epoch 16/20
2/2 - 0s - loss: 0.0620 - val_loss: 0.1009 - 41ms/epoch - 20ms/step
Epoch 17/20
2/2 - 0s - loss: 0.0599 - val_loss: 0.0970 - 44ms/epoch - 22ms/step
Epoch 18/20
2/2 - 0s - loss: 0.0576 - val_loss: 0.0912 - 44ms/epoch - 22ms/step
Epoch 19/20
2/2 - 0s - loss: 0.0553 - val_loss: 0.0848 - 39ms/epoch - 20ms/step
Epoch 20/20
2/2 - 0s - loss: 0.0534 - val_loss: 0.0790 - 39ms/epoch - 20ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.281
AS
Epoch 1/20
2/2 - 3s - loss: 0.2586 - val_loss: 0.2728 - 3s/epoch - 1s/step
Epoch 2/20
2/2 - 0s - loss: 0.1418 - val_loss: 0.1784 - 41ms/epoch - 21ms/step
Epoch 3/20
2/2 - 0s - loss: 0.1053 - val_loss: 0.1295 - 40ms/epoch - 20ms/step
Epoch 4/20
2/2 - 0s - loss: 0.1237 - val_loss: 0.1106 - 42ms/epoch - 21ms/step
Epoch 5/20
2/2 - 0s - loss: 0.1560 - val_loss: 0.1047 - 41ms/epoch - 21ms/step
Epoch 6/20
2/2 - 0s - loss: 0.1717 - val_loss: 0.1025 - 39ms/epoch - 20ms/step
Epoch 7/20
2/2 - 0s - loss: 0.1657 - val_loss: 0.1024 - 39ms/epoch - 20ms/step
Epoch 8/20
2/2 - 0s - loss: 0.1469 - val_loss: 0.1053 - 41ms/epoch - 20ms/step
Epoch 9/20
2/2 - 0s - loss: 0.1252 - val_loss: 0.1113 - 39ms/epoch - 19ms/step
Epoch 10/20
2/2 - 0s - loss: 0.1073 - val_loss: 0.1191 - 41ms/epoch - 20ms/step
Epoch 11/20
2/2 - 0s - loss: 0.0954 - val_loss: 0.1263 - 42ms/epoch - 21ms/step
Epoch 12/20
2/2 - 0s - loss: 0.0885 - val_loss: 0.1310 - 40ms/epoch - 20ms/step
Epoch 13/20
2/2 - 0s - loss: 0.0844 - val_loss: 0.1318 - 40ms/epoch - 20ms/step
Epoch 14/20
2/2 - 0s - loss: 0.0812 - val_loss: 0.1288 - 42ms/epoch - 21ms/step
Epoch 15/20
2/2 - 0s - loss: 0.0781 - val_loss: 0.1230 - 40ms/epoch - 20ms/step
Epoch 16/20
2/2 - 0s - loss: 0.0751 - val_loss: 0.1159 - 38ms/epoch - 19ms/step
Epoch 17/20
2/2 - 0s - loss: 0.0726 - val_loss: 0.1089 - 40ms/epoch - 20ms/step
Epoch 18/20
2/2 - 0s - loss: 0.0710 - val_loss: 0.1031 - 40ms/epoch - 20ms/step
Epoch 19/20
2/2 - 0s - loss: 0.0700 - val_loss: 0.0989 - 40ms/epoch - 20ms/step
Epoch 20/20
2/2 - 0s - loss: 0.0689 - val_loss: 0.0964 - 43ms/epoch - 21ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.311
B6
Epoch 1/20
2/2 - 2s - loss: 0.2732 - val_loss: 0.5680 - 2s/epoch - 1s/step
Epoch 2/20
2/2 - 0s - loss: 0.1925 - val_loss: 0.4114 - 37ms/epoch - 19ms/step
Epoch 3/20
2/2 - 0s - loss: 0.1403 - val_loss: 0.2876 - 37ms/epoch - 18ms/step
Epoch 4/20
2/2 - 0s - loss: 0.1115 - val_loss: 0.1963 - 38ms/epoch - 19ms/step
Epoch 5/20
2/2 - 0s - loss: 0.1023 - val_loss: 0.1341 - 39ms/epoch - 20ms/step
Epoch 6/20
2/2 - 0s - loss: 0.1074 - val_loss: 0.0959 - 40ms/epoch - 20ms/step
Epoch 7/20
2/2 - 0s - loss: 0.1204 - val_loss: 0.0754 - 52ms/epoch - 26ms/step
Epoch 8/20
2/2 - 0s - loss: 0.1348 - val_loss: 0.0663 - 38ms/epoch - 19ms/step
Epoch 9/20
2/2 - 0s - loss: 0.1461 - val_loss: 0.0633 - 38ms/epoch - 19ms/step
Epoch 10/20
2/2 - 0s - loss: 0.1515 - val_loss: 0.0629 - 38ms/epoch - 19ms/step
Epoch 11/20
2/2 - 0s - loss: 0.1505 - val_loss: 0.0635 - 38ms/epoch - 19ms/step
Epoch 12/20
2/2 - 0s - loss: 0.1441 - val_loss: 0.0644 - 39ms/epoch - 19ms/step
Epoch 13/20
2/2 - 0s - loss: 0.1339 - val_loss: 0.0658 - 39ms/epoch - 20ms/step
Epoch 14/20
2/2 - 0s - loss: 0.1218 - val_loss: 0.0679 - 40ms/epoch - 20ms/step
Epoch 15/20
2/2 - 0s - loss: 0.1095 - val_loss: 0.0707 - 38ms/epoch - 19ms/step
Epoch 16/20
2/2 - 0s - loss: 0.0981 - val_loss: 0.0743 - 40ms/epoch - 20ms/step
Epoch 17/20
2/2 - 0s - loss: 0.0882 - val_loss: 0.0782 - 38ms/epoch - 19ms/step
Epoch 18/20
2/2 - 0s - loss: 0.0802 - val_loss: 0.0820 - 39ms/epoch - 19ms/step
Epoch 19/20
2/2 - 0s - loss: 0.0739 - val_loss: 0.0853 - 39ms/epoch - 20ms/step
Epoch 20/20
2/2 - 0s - loss: 0.0691 - val_loss: 0.0878 - 41ms/epoch - 20ms/step
4/4 [==============================] - 0s 3ms/step

Test RMSE: 0.296
DL
Epoch 1/20
1/1 - 3s - loss: 0.5279 - val_loss: 0.5600 - 3s/epoch - 3s/step
Epoch 2/20
1/1 - 0s - loss: 0.4563 - val_loss: 0.4856 - 36ms/epoch - 36ms/step
Epoch 3/20
1/1 - 0s - loss: 0.3912 - val_loss: 0.4180 - 35ms/epoch - 35ms/step
Epoch 4/20
1/1 - 0s - loss: 0.3324 - val_loss: 0.3570 - 36ms/epoch - 36ms/step
Epoch 5/20
1/1 - 0s - loss: 0.2800 - val_loss: 0.3024 - 36ms/epoch - 36ms/step
Epoch 6/20
1/1 - 0s - loss: 0.2336 - val_loss: 0.2541 - 37ms/epoch - 37ms/step
Epoch 7/20
1/1 - 0s - loss: 0.1931 - val_loss: 0.2120 - 38ms/epoch - 38ms/step
Epoch 8/20
1/1 - 0s - loss: 0.1585 - val_loss: 0.1761 - 35ms/epoch - 35ms/step
Epoch 9/20
1/1 - 0s - loss: 0.1296 - val_loss: 0.1462 - 36ms/epoch - 36ms/step
Epoch 10/20
1/1 - 0s - loss: 0.1062 - val_loss: 0.1223 - 36ms/epoch - 36ms/step
Epoch 11/20
1/1 - 0s - loss: 0.0884 - val_loss: 0.1041 - 45ms/epoch - 45ms/step
Epoch 12/20
1/1 - 0s - loss: 0.0756 - val_loss: 0.0911 - 37ms/epoch - 37ms/step
Epoch 13/20
1/1 - 0s - loss: 0.0677 - val_loss: 0.0829 - 37ms/epoch - 37ms/step
Epoch 14/20
1/1 - 0s - loss: 0.0640 - val_loss: 0.0785 - 36ms/epoch - 36ms/step
Epoch 15/20
1/1 - 0s - loss: 0.0637 - val_loss: 0.0771 - 36ms/epoch - 36ms/step
Epoch 16/20
1/1 - 0s - loss: 0.0659 - val_loss: 0.0776 - 36ms/epoch - 36ms/step
Epoch 17/20
1/1 - 0s - loss: 0.0696 - val_loss: 0.0789 - 37ms/epoch - 37ms/step
Epoch 18/20
1/1 - 0s - loss: 0.0736 - val_loss: 0.0802 - 40ms/epoch - 40ms/step
Epoch 19/20
1/1 - 0s - loss: 0.0772 - val_loss: 0.0809 - 38ms/epoch - 38ms/step
Epoch 20/20
1/1 - 0s - loss: 0.0798 - val_loss: 0.0808 - 36ms/epoch - 36ms/step
4/4 [==============================] - 0s 3ms/step

Test RMSE: 0.284
G4
Epoch 1/20
1/1 - 2s - loss: 0.0279 - val_loss: 0.2764 - 2s/epoch - 2s/step
Epoch 2/20
1/1 - 0s - loss: 0.0202 - val_loss: 0.3053 - 36ms/epoch - 36ms/step
Epoch 3/20
1/1 - 0s - loss: 0.0194 - val_loss: 0.3192 - 35ms/epoch - 35ms/step
Epoch 4/20
1/1 - 0s - loss: 0.0182 - val_loss: 0.3206 - 35ms/epoch - 35ms/step
Epoch 5/20
1/1 - 0s - loss: 0.0156 - val_loss: 0.3151 - 35ms/epoch - 35ms/step
Epoch 6/20
1/1 - 0s - loss: 0.0128 - val_loss: 0.3072 - 32ms/epoch - 32ms/step
Epoch 7/20
1/1 - 0s - loss: 0.0108 - val_loss: 0.3007 - 34ms/epoch - 34ms/step
Epoch 8/20
1/1 - 0s - loss: 0.0097 - val_loss: 0.2978 - 32ms/epoch - 32ms/step
Epoch 9/20
1/1 - 0s - loss: 0.0093 - val_loss: 0.2995 - 32ms/epoch - 32ms/step
Epoch 10/20
1/1 - 0s - loss: 0.0087 - val_loss: 0.3055 - 34ms/epoch - 34ms/step
Epoch 11/20
1/1 - 0s - loss: 0.0078 - val_loss: 0.3150 - 32ms/epoch - 32ms/step
Epoch 12/20
1/1 - 0s - loss: 0.0068 - val_loss: 0.3268 - 33ms/epoch - 33ms/step
Epoch 13/20
1/1 - 0s - loss: 0.0059 - val_loss: 0.3389 - 33ms/epoch - 33ms/step
Epoch 14/20
1/1 - 0s - loss: 0.0054 - val_loss: 0.3494 - 33ms/epoch - 33ms/step
Epoch 15/20
1/1 - 0s - loss: 0.0052 - val_loss: 0.3565 - 36ms/epoch - 36ms/step
Epoch 16/20
1/1 - 0s - loss: 0.0050 - val_loss: 0.3596 - 31ms/epoch - 31ms/step
Epoch 17/20
1/1 - 0s - loss: 0.0047 - val_loss: 0.3587 - 37ms/epoch - 37ms/step
Epoch 18/20
1/1 - 0s - loss: 0.0042 - val_loss: 0.3549 - 35ms/epoch - 35ms/step
Epoch 19/20
1/1 - 0s - loss: 0.0037 - val_loss: 0.3497 - 34ms/epoch - 34ms/step
Epoch 20/20
1/1 - 0s - loss: 0.0032 - val_loss: 0.3446 - 33ms/epoch - 33ms/step
2/2 [==============================] - 0s 4ms/step

Test RMSE: 0.587
HA
Epoch 1/20
1/1 - 3s - loss: 0.3407 - val_loss: 0.4060 - 3s/epoch - 3s/step
Epoch 2/20
1/1 - 0s - loss: 0.2746 - val_loss: 0.3141 - 33ms/epoch - 33ms/step
Epoch 3/20
1/1 - 0s - loss: 0.2220 - val_loss: 0.2403 - 32ms/epoch - 32ms/step
Epoch 4/20
1/1 - 0s - loss: 0.1827 - val_loss: 0.1836 - 35ms/epoch - 35ms/step
Epoch 5/20
1/1 - 0s - loss: 0.1558 - val_loss: 0.1426 - 34ms/epoch - 34ms/step
Epoch 6/20
1/1 - 0s - loss: 0.1401 - val_loss: 0.1153 - 36ms/epoch - 36ms/step
Epoch 7/20
1/1 - 0s - loss: 0.1335 - val_loss: 0.0990 - 34ms/epoch - 34ms/step
Epoch 8/20
1/1 - 0s - loss: 0.1332 - val_loss: 0.0905 - 35ms/epoch - 35ms/step
Epoch 9/20
1/1 - 0s - loss: 0.1361 - val_loss: 0.0868 - 34ms/epoch - 34ms/step
Epoch 10/20
1/1 - 0s - loss: 0.1396 - val_loss: 0.0858 - 33ms/epoch - 33ms/step
Epoch 11/20
1/1 - 0s - loss: 0.1415 - val_loss: 0.0863 - 35ms/epoch - 35ms/step
Epoch 12/20
1/1 - 0s - loss: 0.1410 - val_loss: 0.0882 - 36ms/epoch - 36ms/step
Epoch 13/20
1/1 - 0s - loss: 0.1379 - val_loss: 0.0917 - 33ms/epoch - 33ms/step
Epoch 14/20
1/1 - 0s - loss: 0.1325 - val_loss: 0.0974 - 35ms/epoch - 35ms/step
Epoch 15/20
1/1 - 0s - loss: 0.1257 - val_loss: 0.1057 - 35ms/epoch - 35ms/step
Epoch 16/20
1/1 - 0s - loss: 0.1182 - val_loss: 0.1171 - 35ms/epoch - 35ms/step
Epoch 17/20
1/1 - 0s - loss: 0.1108 - val_loss: 0.1315 - 34ms/epoch - 34ms/step
Epoch 18/20
1/1 - 0s - loss: 0.1040 - val_loss: 0.1489 - 33ms/epoch - 33ms/step
Epoch 19/20
1/1 - 0s - loss: 0.0982 - val_loss: 0.1689 - 33ms/epoch - 33ms/step
Epoch 20/20
1/1 - 0s - loss: 0.0936 - val_loss: 0.1908 - 33ms/epoch - 33ms/step
1/1 [==============================] - 0s 407ms/step

Test RMSE: 0.437
NK
Epoch 1/20
1/1 - 3s - loss: 0.5218 - val_loss: 0.2882 - 3s/epoch - 3s/step
Epoch 2/20
1/1 - 0s - loss: 0.4567 - val_loss: 0.2644 - 36ms/epoch - 36ms/step
Epoch 3/20
1/1 - 0s - loss: 0.3965 - val_loss: 0.2497 - 41ms/epoch - 41ms/step
Epoch 4/20
1/1 - 0s - loss: 0.3414 - val_loss: 0.2446 - 40ms/epoch - 40ms/step
Epoch 5/20
1/1 - 0s - loss: 0.2916 - val_loss: 0.2491 - 38ms/epoch - 38ms/step
Epoch 6/20
1/1 - 0s - loss: 0.2472 - val_loss: 0.2632 - 39ms/epoch - 39ms/step
Epoch 7/20
1/1 - 0s - loss: 0.2084 - val_loss: 0.2863 - 40ms/epoch - 40ms/step
Epoch 8/20
1/1 - 0s - loss: 0.1754 - val_loss: 0.3172 - 40ms/epoch - 40ms/step
Epoch 9/20
1/1 - 0s - loss: 0.1481 - val_loss: 0.3543 - 35ms/epoch - 35ms/step
Epoch 10/20
1/1 - 0s - loss: 0.1264 - val_loss: 0.3950 - 37ms/epoch - 37ms/step
Epoch 11/20
1/1 - 0s - loss: 0.1100 - val_loss: 0.4367 - 35ms/epoch - 35ms/step
Epoch 12/20
1/1 - 0s - loss: 0.0985 - val_loss: 0.4763 - 36ms/epoch - 36ms/step
Epoch 13/20
1/1 - 0s - loss: 0.0911 - val_loss: 0.5111 - 38ms/epoch - 38ms/step
Epoch 14/20
1/1 - 0s - loss: 0.0869 - val_loss: 0.5389 - 35ms/epoch - 35ms/step
Epoch 15/20
1/1 - 0s - loss: 0.0850 - val_loss: 0.5579 - 36ms/epoch - 36ms/step
Epoch 16/20
1/1 - 0s - loss: 0.0845 - val_loss: 0.5675 - 36ms/epoch - 36ms/step
Epoch 17/20
1/1 - 0s - loss: 0.0844 - val_loss: 0.5676 - 36ms/epoch - 36ms/step
Epoch 18/20
1/1 - 0s - loss: 0.0841 - val_loss: 0.5591 - 35ms/epoch - 35ms/step
Epoch 19/20
1/1 - 0s - loss: 0.0831 - val_loss: 0.5429 - 35ms/epoch - 35ms/step
Epoch 20/20
1/1 - 0s - loss: 0.0812 - val_loss: 0.5206 - 35ms/epoch - 35ms/step
3/3 [==============================] - 0s 4ms/step

Test RMSE: 0.722
UA
Epoch 1/20
2/2 - 3s - loss: 0.3601 - val_loss: 0.4134 - 3s/epoch - 1s/step
Epoch 2/20
2/2 - 0s - loss: 0.2574 - val_loss: 0.2877 - 39ms/epoch - 19ms/step
Epoch 3/20
2/2 - 0s - loss: 0.2029 - val_loss: 0.1943 - 38ms/epoch - 19ms/step
Epoch 4/20
2/2 - 0s - loss: 0.1885 - val_loss: 0.1330 - 40ms/epoch - 20ms/step
Epoch 5/20
2/2 - 0s - loss: 0.2026 - val_loss: 0.0981 - 38ms/epoch - 19ms/step
Epoch 6/20
2/2 - 0s - loss: 0.2273 - val_loss: 0.0808 - 40ms/epoch - 20ms/step
Epoch 7/20
2/2 - 0s - loss: 0.2465 - val_loss: 0.0735 - 40ms/epoch - 20ms/step
Epoch 8/20
2/2 - 0s - loss: 0.2519 - val_loss: 0.0718 - 42ms/epoch - 21ms/step
Epoch 9/20
2/2 - 0s - loss: 0.2436 - val_loss: 0.0737 - 38ms/epoch - 19ms/step
Epoch 10/20
2/2 - 0s - loss: 0.2265 - val_loss: 0.0786 - 40ms/epoch - 20ms/step
Epoch 11/20
2/2 - 0s - loss: 0.2062 - val_loss: 0.0860 - 40ms/epoch - 20ms/step
Epoch 12/20
2/2 - 0s - loss: 0.1869 - val_loss: 0.0945 - 42ms/epoch - 21ms/step
Epoch 13/20
2/2 - 0s - loss: 0.1708 - val_loss: 0.1027 - 39ms/epoch - 20ms/step
Epoch 14/20
2/2 - 0s - loss: 0.1583 - val_loss: 0.1092 - 42ms/epoch - 21ms/step
Epoch 15/20
2/2 - 0s - loss: 0.1489 - val_loss: 0.1128 - 40ms/epoch - 20ms/step
Epoch 16/20
2/2 - 0s - loss: 0.1417 - val_loss: 0.1131 - 39ms/epoch - 20ms/step
Epoch 17/20
2/2 - 0s - loss: 0.1359 - val_loss: 0.1104 - 41ms/epoch - 20ms/step
Epoch 18/20
2/2 - 0s - loss: 0.1311 - val_loss: 0.1052 - 41ms/epoch - 20ms/step
Epoch 19/20
2/2 - 0s - loss: 0.1272 - val_loss: 0.0986 - 40ms/epoch - 20ms/step
Epoch 20/20
2/2 - 0s - loss: 0.1240 - val_loss: 0.0915 - 39ms/epoch - 20ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.302
WN
Epoch 1/20
2/2 - 2s - loss: 0.2861 - val_loss: 0.1369 - 2s/epoch - 1s/step
Epoch 2/20
2/2 - 0s - loss: 0.1983 - val_loss: 0.1659 - 42ms/epoch - 21ms/step
Epoch 3/20
2/2 - 0s - loss: 0.1582 - val_loss: 0.2075 - 47ms/epoch - 24ms/step
Epoch 4/20
2/2 - 0s - loss: 0.1468 - val_loss: 0.2316 - 39ms/epoch - 20ms/step
Epoch 5/20
2/2 - 0s - loss: 0.1440 - val_loss: 0.2308 - 43ms/epoch - 22ms/step
Epoch 6/20
2/2 - 0s - loss: 0.1398 - val_loss: 0.2135 - 40ms/epoch - 20ms/step
Epoch 7/20
2/2 - 0s - loss: 0.1335 - val_loss: 0.1895 - 41ms/epoch - 21ms/step
Epoch 8/20
2/2 - 0s - loss: 0.1277 - val_loss: 0.1660 - 40ms/epoch - 20ms/step
Epoch 9/20
2/2 - 0s - loss: 0.1241 - val_loss: 0.1470 - 41ms/epoch - 20ms/step
Epoch 10/20
2/2 - 0s - loss: 0.1227 - val_loss: 0.1336 - 38ms/epoch - 19ms/step
Epoch 11/20
2/2 - 0s - loss: 0.1222 - val_loss: 0.1252 - 40ms/epoch - 20ms/step
Epoch 12/20
2/2 - 0s - loss: 0.1211 - val_loss: 0.1210 - 39ms/epoch - 19ms/step
Epoch 13/20
2/2 - 0s - loss: 0.1183 - val_loss: 0.1200 - 53ms/epoch - 27ms/step
Epoch 14/20
2/2 - 0s - loss: 0.1138 - val_loss: 0.1218 - 50ms/epoch - 25ms/step
Epoch 15/20
2/2 - 0s - loss: 0.1087 - val_loss: 0.1255 - 40ms/epoch - 20ms/step
Epoch 16/20
2/2 - 0s - loss: 0.1037 - val_loss: 0.1301 - 39ms/epoch - 19ms/step
Epoch 17/20
2/2 - 0s - loss: 0.0994 - val_loss: 0.1340 - 39ms/epoch - 20ms/step
Epoch 18/20
2/2 - 0s - loss: 0.0958 - val_loss: 0.1359 - 39ms/epoch - 20ms/step
Epoch 19/20
2/2 - 0s - loss: 0.0926 - val_loss: 0.1350 - 39ms/epoch - 20ms/step
Epoch 20/20
2/2 - 0s - loss: 0.0892 - val_loss: 0.1313 - 45ms/epoch - 23ms/step
5/5 [==============================] - 0s 3ms/step

Test RMSE: 0.362

lstm_results50 = pd.DataFrame.from_dict(lstm_rsme_dict50,  orient='index',
                       columns=['RMSE', 'Epochs', 'Batch Size'])
lstm_results50 = lstm_results50.reset_index()

lstm_results50['Airline'] = lstm_results50['index'].map(MKT_code)
lstm_results50.drop('index', axis=1, inplace=True)
lstm_results50 = lstm_results50.set_index('Airline')
lstm_results50

lstm_results20 = pd.DataFrame.from_dict(lstm_rsme_dict20,  orient='index',
                       columns=['RMSE', 'Epochs', 'Batch Size'])
lstm_results20 = lstm_results20.reset_index()

lstm_results20['Airline'] = lstm_results20['index'].map(MKT_code)
lstm_results20.drop('index', axis=1, inplace=True)
lstm_results20 = lstm_results20.set_index('Airline')
lstm_results20

	Open	High	Low	Close	Adj Close	Volume
Date
2015-04-30	24.493999	24.579000	24.129999	24.188000	23.045496	7300
2015-05-01	24.700001	24.969000	24.379999	24.870001	23.695282	182800
2015-05-04	25.100000	25.100000	24.600000	24.660000	23.495201	209000
2015-05-05	24.480000	24.480000	24.000000	24.000000	22.866375	75400
2015-05-06	24.010000	24.110001	23.590000	23.980000	22.847322	36900

	United	Alaskan	American	Allegiant	JetBlue	Hawaiian	Southwest	Spirit	Delta	Frontier	JETS
Date
2021-04-05	0.025424	0.023025	0.015507	0.021900	0.039273	0.005968	0.027243	0.003480	0.029080	0.014854	0.020702
2021-04-06	0.003205	0.037277	-0.007016	0.007277	0.003779	0.002596	0.017945	0.021339	0.027861	0.031887	0.008330
2021-04-07	0.001513	-0.018714	-0.005403	-0.005557	-0.009412	-0.026257	-0.009048	-0.019849	-0.028074	0.042553	-0.009698
2021-04-08	-0.002350	0.004561	-0.011701	0.001916	0.001900	-0.003038	-0.000945	0.003730	-0.007769	0.032556	-0.002902
2021-04-09	-0.016827	-0.000825	-0.004651	0.007211	-0.001423	-0.006476	-0.000630	-0.011680	-0.010841	0.005647	-0.006912

	Oil Return	United	Alaskan	American	Allegiant	JetBlue	Hawaiian	Southwest	Spirit	Delta	Frontier	JETS
Date
2021-04-05	-0.045565	0.025424	0.023025	0.015507	0.021900	0.039273	0.005968	0.027243	0.003480	0.029080	0.014854	0.020702
2021-04-06	0.011594	0.003205	0.037277	-0.007016	0.007277	0.003779	0.002596	0.017945	0.021339	0.027861	0.031887	0.008330
2021-04-07	0.007416	0.001513	-0.018714	-0.005403	-0.005557	-0.009412	-0.026257	-0.009048	-0.019849	-0.028074	0.042553	-0.009698
2021-04-08	-0.002844	-0.002350	0.004561	-0.011701	0.001916	0.001900	-0.003038	-0.000945	0.003730	-0.007769	0.032556	-0.002902
2021-04-09	-0.004698	-0.016827	-0.000825	-0.004651	0.007211	-0.001423	-0.006476	-0.000630	-0.011680	-0.010841	0.005647	-0.006912

	7	14	30	60	90	180	365
United	0.008827	0.001552	-0.012239	0.025134	-0.063869	0.002089	-0.03858
Alaskan	-0.023877	-0.000322	0.021835	-0.010173	-0.05357	0.036831	-0.015939
American	-0.017728	0.001487	0.006128	0.044039	-0.058847	0.016144	-0.029939
Allegiant	-0.010919	-0.022085	0.015095	0.049887	-0.068738	0.003521	-0.03009
JetBlue	-0.023745	0.018836	0.042487	0.029154	-0.052595	0.075384	-0.042281
Hawaiian	-0.026035	0.021067	0.001887	0.019371	0.009924	-0.023589	-0.073234
Southwest	-0.031236	-0.004643	0.020085	0.010116	0.002631	0.000623	-0.040111
Spirit	0.027841	0.004721	0.003411	0.003366	-0.008387	0.056605	-0.043658
Delta	0.000773	-0.013916	0.00983	0.015437	-0.042934	0.012249	-0.018705
JETS	-0.021892	0.001033	0.005418	0.02165	-0.051936	0.026966	-0.056063

	DL_NUM_DEP_DEL15	DL_NUM_ARR_DEL15	DL_NUM_FL	DL_NUM_CAN	FL_PROP_OP_BY_MKT_DL	CAN_PROP_OP_BY_MKT_DL	DL_PROP_FULL_SEATS	DL_PROP_FREIGHT/MAIL	CANCEL_CODE_CARRIER	CANCEL_CODE_WEATHER	...	Price Change	Trend	Prev Trend	Crude Oil Open	Crude Oil High	Crude Oil Low	Crude Oil Close	Crude Oil Adj Close	Crude Oil Volume	Crude Oil Daily Return
DATE
2021-11-01	417.0	398.0	4167	6.0	0.572354	0.0	0.800704	0.008555	6.0	0.0	...	1.289997	up	down	83.360001	84.879997	82.739998	84.050003	84.050003	412309.0	0.005744
2021-11-02	244.0	233.0	3812	NaN	0.577650	0.0	0.800704	0.008555	NaN	NaN	...	-0.029999	down	up	83.870003	84.410004	82.919998	83.910004	83.910004	429608.0	-0.001666
2021-11-03	300.0	341.0	3837	NaN	0.577013	0.0	0.800704	0.008555	NaN	NaN	...	0.850002	up	down	83.070000	83.080002	79.690002	80.860001	80.860001	546577.0	-0.036349
2021-11-04	396.0	511.0	4247	NaN	0.565576	0.0	0.800704	0.008555	NaN	NaN	...	-0.250000	down	up	80.180000	83.419998	78.250000	78.809998	78.809998	701119.0	-0.025352
2021-11-05	439.0	467.0	4273	NaN	0.564241	0.0	0.800704	0.008555	NaN	NaN	...	3.299999	up	down	79.370003	81.800003	78.959999	81.269997	81.269997	531054.0	0.031214

Stock data ETL¶

Correlational Analysis¶

Load in Data about operational metrics¶

Stock Price Trend Transition Probabilities¶

Process Overview¶

Transition Matrix¶

Marketing Carrier On-Time Performance Data¶

Marketing Carrier On-Time Performance Data ETL¶

Flight Data EDA¶

Air Carrier Statistics (Form 41 Traffic)- U.S. Carriers¶

T-100 Domestic Segment (U.S. Carriers) Data ETL¶

Combining All Data Sets by the Market Airline¶

Airline Operations Performance Metrics¶

Stock and Airline Operations Performance Metrics¶

Combing all Stock and Airline Operations Performance Metrics to one DataFrame¶

Creating a TSNE¶

TSNE using only stock and finacial data¶

TNSE using all data¶

Models¶

Logistic Regression¶

Multivariate LSTM¶

Conclusion¶

	NET_INCOME	OP_PROFIT_LOSS	TRANS_REV_PAX	MAIL	TOTAL_PROPERTY	PROP_FREIGHT	PROP_BAG	TOTAL_CHARTER	CHARTER_PAX	CHARTER_PROP	...	UNIQUE_CARRIER	UNIQUE_CARRIER_NAME	CARRIER	CARRIER_NAME	UNIQUE_CARRIER_ENTITY	REGION	CARRIER_GROUP_NEW	CARRIER_GROUP	YEAR	QUARTER
408	262685.0	-597883.0	NaN	726.0	NaN	444323.0	NaN	NaN	NaN	7892.0	...	FX	Federal Express Corporation	FX	Federal Express Corporation	16196	A	3	3	2020	3
409	688988.0	1395899.0	NaN	0.0	NaN	2785917.0	NaN	NaN	NaN	7598.0	...	FX	Federal Express Corporation	FX	Federal Express Corporation	06200	D	3	3	2020	3
410	908014.0	1130748.0	NaN	394.0	NaN	2566955.0	NaN	NaN	NaN	1195.0	...	FX	Federal Express Corporation	FX	Federal Express Corporation	06200	D	3	3	2020	1
411	926627.0	1363349.0	NaN	302.0	NaN	2502021.0	NaN	NaN	NaN	33790.0	...	FX	Federal Express Corporation	FX	Federal Express Corporation	06200	D	3	3	2020	2
412	1188167.0	1624892.0	NaN	45.0	NaN	2986833.0	NaN	NaN	NaN	6047.0	...	FX	Federal Express Corporation	FX	Federal Express Corporation	06200	D	3	3	2020	4

	NET_INCOME	OP_PROFIT_LOSS	TRANS_REV_PAX	MAIL	TOTAL_PROPERTY	PROP_FREIGHT	PROP_BAG	TOTAL_CHARTER	CHARTER_PAX	CHARTER_PROP	...	UNIQUE_CARRIER	UNIQUE_CARRIER_NAME	CARRIER	CARRIER_NAME	UNIQUE_CARRIER_ENTITY	REGION	CARRIER_GROUP_NEW	CARRIER_GROUP	YEAR	QUARTER
2	-294047.62	-294047.62	895106.59	7515.70	NaN	124382.84	8503.14	NaN	0.00	0.0	...	AA	American Airlines Inc.	AA	American Airlines Inc.	10049	A	3	3	2023	1
3	-258801.65	-234102.19	5015655.46	10224.02	NaN	68087.22	210466.52	NaN	17355.33	NaN	...	UA	United Air Lines Inc.	UA	United Air Lines Inc.	0A875	D	3	3	2023	1
4	-233930.65	-295070.82	1464226.29	0.00	NaN	101.75	104368.02	NaN	224.13	0.0	...	B6	JetBlue Airways	B6	JetBlue Airways	06673	D	3	3	2023	1
5	-185853.00	437900.00	5914695.00	9697.00	NaN	60430.00	172908.00	NaN	77822.00	85.0	...	DL	Delta Air Lines Inc.	DL	Delta Air Lines Inc.	01260	D	3	3	2023	1
6	-153850.00	-274205.00	4697111.00	587.00	NaN	38321.00	15128.00	NaN	5787.00	0.0	...	WN	Southwest Airlines Co.	WN	Southwest Airlines Co.	06725	D	3	3	2023	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1336	6542.00	22237.00	637215.00	0.00	NaN	9743.00	26125.00	NaN	3724.00	0.0	...	DL	Delta Air Lines Inc.	DL	Delta Air Lines Inc.	10260	L	3	3	2020	1
1350	9127.98	14530.18	30357.19	0.00	NaN	0.00	4681.78	NaN	0.00	0.0	...	F9	Frontier Airlines Inc.	F9	Frontier Airlines Inc.	16461	L	3	3	2020	1
1356	11141.00	30232.00	744195.00	5898.00	NaN	58152.00	16292.00	NaN	8705.00	0.0	...	DL	Delta Air Lines Inc.	DL	Delta Air Lines Inc.	10261	A	3	3	2020	1
1366	17296.17	-80572.10	130274.24	0.00	NaN	0.00	17308.07	NaN	0.00	0.0	...	F9	Frontier Airlines Inc.	F9	Frontier Airlines Inc.	06460	D	3	3	2020	2
1384	29778.84	38659.28	299862.94	0.00	NaN	0.00	57336.92	NaN	8918.84	0.0	...	G4	Allegiant Air	G4	Allegiant Air	06035	D	3	3	2020	1

	UNIQUE_CARRIER	YEAR	QUARTER	OP_PROFIT_LOSS
0	AA	2019	4	-2535836.17
1	AA	2020	1	-4459137.82
2	AA	2020	2	-4945286.70
3	AA	2020	3	-2438663.13
4	AA	2020	4	-3439576.16
...	...	...	...	...
135	WN	2022	1	1158160.00
136	WN	2022	2	395148.00
137	WN	2022	3	-385541.00
138	WN	2022	4	-284229.00
139	WN	2023	1	794871.00

	MKT_UNIQUE_CARRIER	YEAR	QUARTER	MONTH	DAY_OF_MONTH	DAY_OF_WEEK	FL_DATE	MKT_CARRIER	OP_UNIQUE_CARRIER	OP_CARRIER	...	ARR_DELAY	ARR_DELAY_NEW	ARR_DELAY_GROUP
0	AA	2023	2	5	16	2	2023-05-16	AA	AA	AA	...	-32.0	0.0	-2.0
1	AA	2023	2	5	16	2	2023-05-16	AA	AA	AA	...	4.0	4.0	0.0
2	AA	2023	2	5	16	2	2023-05-16	AA	AA	AA	...	12.0	12.0	0.0
3	AA	2023	2	5	16	2	2023-05-16	AA	AA	AA	...	-9.0	0.0	-1.0
4	AA	2023	2	5	16	2	2023-05-16	AA	AA	AA	...	-28.0	0.0	-2.0

	UNIQUE_CARRIER	DISTANCE	RAMP_TO_RAMP	...	DEST_STATE_NM	DEST_WAC	AIRCRAFT_GROUP	AIRCRAFT_TYPE	AIRCRAFT_CONFIG	YEAR	QUARTER	MONTH	DISTANCE_GROUP	CLASS
0	YV	24.0	32.0	...	Virginia	38	6	673	1	2022	3	7	1	F
1	YV	45.0	35.0	...	Virginia	38	6	673	1	2022	3	7	1	F
2	YV	70.0	52.0	...	Pennsylvania	23	6	673	1	2022	1	3	1	F
3	YV	83.0	32.0	...	Florida	33	6	673	1	2022	2	5	1	F
4	YV	83.0	52.0	...	North Carolina	36	6	673	1	2022	2	6	1	F

	Date	NUM_DEP_DEL15	NUM_ARR_DEL15	NUM_FL	NUM_CAN	FL_PROP_OP_BY_MKT	CAN_PROP_OP_BY_MKT	PROP_FULL_SEATS	PROP_FREIGHT/MAIL	CANCEL_CODE_CARRIER	...	Trend	Prev Trend	Crude Oil Open	Crude Oil High	Crude Oil Low	Crude Oil Close	Crude Oil Adj Close	Crude Oil Volume	Crude Oil Daily Return	Prev Quarter OP_PROFIT_LOSS
0	2021-11-01	845.0	791.0	3756	75.0	0.432109	0.146667	0.800704	0.008555	56.0	...	1.0	1.0	83.360001	84.879997	82.739998	84.050003	84.050003	412309.0	0.005744	-1375235.78
1	2021-11-02	507.0	478.0	3688	44.0	0.428688	0.181818	0.800704	0.008555	37.0	...	0.0	1.0	83.870003	84.410004	82.919998	83.910004	83.910004	429608.0	-0.001666	-1375235.78
2	2021-11-03	627.0	718.0	3726	30.0	0.424047	0.300000	0.800704	0.008555	25.0	...	1.0	0.0	83.070000	83.080002	79.690002	80.860001	80.860001	546577.0	-0.036349	-1375235.78
3	2021-11-04	540.0	609.0	3797	11.0	0.429550	0.454545	0.800704	0.008555	7.0	...	1.0	1.0	80.180000	83.419998	78.250000	78.809998	78.809998	701119.0	-0.025352	-1375235.78
4	2021-11-05	420.0	421.0	3750	15.0	0.426667	0.266667	0.800704	0.008555	11.0	...	1.0	1.0	79.370003	81.800003	78.959999	81.269997	81.269997	531054.0	0.031214	-1375235.78

	Random State 42		Random State 104		Random State 35
	Accuracy	MSE	Accuracy	MSE	Accuracy	MSE
Airline
AA	0.987500	0.012500	0.962500	0.037500	0.925000	0.075000
AS	0.948718	0.051282	0.961538	0.038462	0.974359	0.025641
B6	0.971014	0.028986	0.942029	0.057971	0.985507	0.014493
DL	0.925373	0.074627	0.910448	0.089552	0.970149	0.029851
G4	0.958333	0.041667	0.958333	0.041667	0.895833	0.104167
HA	0.966667	0.033333	0.900000	0.100000	0.933333	0.066667
NK	0.949153	0.050847	0.864407	0.135593	0.983051	0.016949
UA	0.912500	0.087500	0.912500	0.087500	0.912500	0.087500
WN	0.950000	0.050000	0.950000	0.050000	0.875000	0.125000

	RMSE	Epochs	Batch Size
Airline
American	0.440847	50	72
Alaskan	0.388948	50	72
JetBlue	0.304574	50	72
Delta	0.214968	50	72
Allegiant	0.555687	50	72
Hawaiian	0.698020	50	72
Spirit	0.395131	50	72
United	0.236709	50	72
Southwest	0.302007	50	72

	RMSE	Epochs	Batch Size
Airline
American	0.280981	20	72
Alaskan	0.310535	20	72
JetBlue	0.296296	20	72
Delta	0.284168	20	72
Allegiant	0.587060	20	72
Hawaiian	0.436785	20	72
Spirit	0.721506	20	72
United	0.302426	20	72
Southwest	0.362422	20	72

	Random State 42		Random State 104		Random State 35
	Accuracy	MSE	Accuracy	MSE	Accuracy	MSE
Airline
AA	0.612500	0.387500	0.550000	0.450000	0.575000	0.425000
AS	0.487179	0.512821	0.500000	0.500000	0.576923	0.423077
B6	0.449275	0.550725	0.507246	0.492754	0.434783	0.565217
DL	0.424242	0.575758	0.409091	0.590909	0.424242	0.575758
G4	0.458333	0.541667	0.520833	0.479167	0.562500	0.437500
HA	0.700000	0.300000	0.666667	0.333333	0.566667	0.433333
NK	0.610169	0.389831	0.491525	0.508475	0.559322	0.440678
UA	0.425000	0.575000	0.525000	0.475000	0.487500	0.512500
WN	0.475000	0.525000	0.500000	0.500000	0.525000	0.475000