Climate Prediction Using LSTM

This project aims to predict the mean temperature in Delhi using an LSTM (Long Short-Term Memory) neural network. The data includes daily climate observations such as temperature, humidity, wind speed, and pressure.

Table of Contents

• Dataset
• Requirements
• Model
• Training and Evaluation
• Results
• License

Dataset

The dataset used in this project is split into two files:

• DailyDelhiClimateTrain.csv
• DailyDelhiClimateTest.csv

Both files contain the following columns:

• meantemp: The mean temperature
• humidity: The humidity level
• wind_speed: The wind speed
• meanpressure: The mean pressure

Requirements

To run this project, you need the following dependencies:

• Python 3.7+
• NumPy
• Pandas
• PyTorch
• Scikit-learn

You can install the dependencies using pip:

pip install numpy pandas torch scikit-learn

Model

The LSTM model is defined in main.ipynb. The key parameters are:

input_size: Number of input features (4 in this case: meantemp, humidity, wind_speed, meanpressure)
hidden_size: Number of hidden units in the LSTM layer (64)
num_layers: Number of LSTM layers (2)
output_size: Number of output features (1, since we are predicting meantemp)

Training and Evaluation

The training process involves the following steps:

Reading and preprocessing the data: The data is read from CSV files and converted to NumPy arrays. Creating sequences: The data is transformed into sequences to fit the LSTM model.

Creating DataLoaders: PyTorch DataLoaders are created for batching the data.

Defining the model: An LSTM model is defined using PyTorch.

Training the model: The model is trained using Mean Squared Error (MSE) loss and Adam optimizer.

Evaluating the model: The model is evaluated on the test set to compute the test loss, Mean Squared Error (MSE), and R² score.

Results

After training, the model achieved the following performance on the test set:

Test Loss: 4.4735 Mean Squared Error: 4.7441 R² Score: 0.8852 Here are the first five actual vs. predicted temperatures:

Actual: [15.684211, 14.571428, 12.111111, 11.0, 11.789474]
Predicted: [28.027569, 15.272582, 14.947316, 14.326501, 13.310998]