load_and_test_ML.py

import os
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf

# Load the pre-trained model
model = tf.keras.models.load_model("model/LSTM_1_model_saved_model")

# Load the test dataset
test = pd.read_csv("dataset/test.csv")

# Ensure the new data is preprocessed in the same way as the training data
scaler = tf.keras.layers.experimental.preprocessing.Normalization()
scaler.adapt(test[['arrival_hour', 'arrival_minute', 'stop_lat', 'stop_lon', 'next_lat', 'next_lon', 'direction_id']])
X_new = scaler(test[['arrival_hour', 'arrival_minute', 'stop_lat', 'stop_lon', 'next_lat', 'next_lon', 'direction_id']])

# Define the sequence length (seq_length) used during training
seq_length = 64

# Generate sequences from the new data
X_seq_new = []
num_objects = len(X_new)

for i in range(num_objects):
    start = max(0, i - seq_length + 1)
    sequence = X_new[start:i + 1]
    X_seq_new.append(sequence)

# Pad sequences to ensure consistent length
X_seq_new = tf.keras.preprocessing.sequence.pad_sequences(X_seq_new, maxlen=seq_length, dtype='float32', padding='post', truncating='post')

# Assuming 'arrival_time' is the column in the 'test' DataFrame representing the time steps
timestamps = test['arrival_time'].iloc[seq_length - 1:num_objects]

# Make predictions
predicted_congestion = model.predict(X_seq_new)

# Create a DataFrame with location and predicted congestion
location = test[['arrival_hour', 'arrival_minute', 'stop_lat', 'stop_lon', 'next_lat', 'next_lon', 'direction_id']].copy()
location['congestion_level'] = predicted_congestion

print(location)

# Plot the predicted and actual congestion levels
plt.figure(figsize=(12, 6))
plt.plot(test['congestion_level'], label="Actual Congestion")
plt.plot(predicted_congestion, label="Predicted Congestion")
plt.legend()
plt.xlabel("Time Steps - Location")
plt.ylabel("Congestion Level")
plt.title("Actual vs. Predicted Congestion Levels")
plt.show()