import numpy as np
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.metrics import r2_score, mean_squared_error
import tensorflow as tf
from tensorflow.keras import layers, models
import seaborn as sns
import pandas as pd

# Load the 'flights' dataset from seaborn
flights = sns.load_dataset('flights')

flights

# Prepare the time series data
time_series_data = flights['passengers'].values
time_steps = 12  # Using a window of 12 months (1 year)
X, y = [], []
for i in range(len(time_series_data) - time_steps):
    X.append(time_series_data[i:i + time_steps])
    y.append(time_series_data[i + time_steps])
X = np.array(X)
y = np.array(y)

# Reshape X for 1D CNN
X = X.reshape(-1, time_steps, 1)
print(time_series_data[:15])
print(X[0,:,0])
print(X[1,:,0])
print(y[:2])
print(X.shape)
print(y.shape)

[112 118 132 129 121 135 148 148 136 119 104 118 115 126 141]
[112 118 132 129 121 135 148 148 136 119 104 118]
[118 132 129 121 135 148 148 136 119 104 118 115]
[115 126]
(132, 12, 1)
(132,)

# Plot the full time series
t = np.arange(len(time_series_data))
plt.figure(figsize=(12, 3))
plt.plot(t, time_series_data, label='Airline Passengers')
plt.xlabel('Time (Months)')
plt.ylabel('Number of Passengers')
plt.title('Airline Passengers Over Time')
plt.legend()
plt.show()

# Split the data into training and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Build the 1D CNN model
model = models.Sequential([
    layers.Conv1D(32, kernel_size=3, activation='relu', input_shape=(time_steps, 1)),
    layers.MaxPooling1D(pool_size=2),
    layers.Conv1D(64, kernel_size=3, activation='relu'),
    layers.GlobalAveragePooling1D(),
    layers.Dense(50, activation='relu'),
    layers.Dense(1)
])

# Compile the model
model.compile(optimizer='adam', loss='mean_squared_error')

# Train the model
history = model.fit(X_train, y_train, epochs=25, validation_split=0.1)

Epoch 1/25

/usr/local/lib/python3.10/dist-packages/keras/src/layers/convolutional/base_conv.py:107: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.
  super().__init__(activity_regularizer=activity_regularizer, **kwargs)

3/3 ━━━━━━━━━━━━━━━━━━━━ 2s 139ms/step - loss: 81642.4531 - val_loss: 78684.9453
Epoch 2/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 35ms/step - loss: 75499.3281 - val_loss: 65682.4219
Epoch 3/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 59268.0625 - val_loss: 53558.5273
Epoch 4/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 47788.9883 - val_loss: 40289.0430
Epoch 5/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 37886.7148 - val_loss: 27205.5801
Epoch 6/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 25ms/step - loss: 25103.9746 - val_loss: 15658.7041
Epoch 7/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 22ms/step - loss: 16134.2773 - val_loss: 7122.2417
Epoch 8/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 41ms/step - loss: 8428.2207 - val_loss: 3933.6521
Epoch 9/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 27ms/step - loss: 3732.4946 - val_loss: 6675.1333
Epoch 10/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 4305.7979 - val_loss: 11658.7129
Epoch 11/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 16ms/step - loss: 6776.3887 - val_loss: 13384.7041
Epoch 12/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 16ms/step - loss: 7420.2119 - val_loss: 10881.7100
Epoch 13/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 17ms/step - loss: 5401.3750 - val_loss: 7423.1079
Epoch 14/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 16ms/step - loss: 3812.3354 - val_loss: 5009.0483
Epoch 15/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 23ms/step - loss: 3596.3770 - val_loss: 4057.4211
Epoch 16/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 19ms/step - loss: 4162.7432 - val_loss: 3884.2405
Epoch 17/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 17ms/step - loss: 3952.0967 - val_loss: 3868.0859
Epoch 18/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 26ms/step - loss: 4106.3691 - val_loss: 3896.0859
Epoch 19/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 21ms/step - loss: 3715.5227 - val_loss: 4101.0923
Epoch 20/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 17ms/step - loss: 3684.1990 - val_loss: 4522.3345
Epoch 21/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 16ms/step - loss: 3683.0820 - val_loss: 5051.2363
Epoch 22/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 17ms/step - loss: 3497.8530 - val_loss: 5381.6528
Epoch 23/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 16ms/step - loss: 3558.4512 - val_loss: 5386.2437
Epoch 24/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 19ms/step - loss: 3438.4111 - val_loss: 5104.2617
Epoch 25/25
3/3 ━━━━━━━━━━━━━━━━━━━━ 0s 16ms/step - loss: 3664.5737 - val_loss: 4751.6230

# Evaluate the model
test_loss = model.evaluate(X_test, y_test)
print(f"Test loss: {test_loss:.4f}")

# Plot training history
plt.figure(figsize=(6,3))
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Loss')
plt.legend(loc='upper right')
plt.show()

# Predict and plot results
predictions = model.predict(X_test)
plt.figure(figsize=(7,3))
plt.plot(y_test, label='True Values')
plt.plot(predictions, label='Predictions')
plt.xlabel('Sample Index')
plt.ylabel('Number of Passengers')
plt.legend()
plt.show()

1/1 ━━━━━━━━━━━━━━━━━━━━ 0s 34ms/step - loss: 2351.5320
Test loss: 2351.5320

1/1 ━━━━━━━━━━━━━━━━━━━━ 0s 142ms/step

# Calculate R-squared and correlation coefficient
r2 = r2_score(y_test, predictions)
correlation_matrix = np.corrcoef(y_test, predictions.flatten())
correlation_coefficient = correlation_matrix[0, 1]

print(f"R-squared: {r2:.4f}")
print(f"Correlation Coefficient: {correlation_coefficient:.4f}")

# Create scatter plot
plt.figure(figsize=(4,4))
plt.scatter(y_test, predictions, alpha=0.6)
plt.xlabel('True Values')
plt.ylabel('Predicted Values')
plt.title('Scatter Plot of True vs. Predicted Values')
plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], 'k--')
plt.grid()
plt.show()

R-squared: 0.8119
Correlation Coefficient: 0.9048

	year	month	passengers
0	1949	Jan	112
1	1949	Feb	118
2	1949	Mar	132
3	1949	Apr	129
4	1949	May	121
...	...	...	...
139	1960	Aug	606
140	1960	Sep	508
141	1960	Oct	461
142	1960	Nov	390
143	1960	Dec	432