Rain Prediction Model

Project Overview

This project builds a machine learning model to predict whether it will rain tomorrow based on historical weather data. It uses features like temperature, humidity, wind speed, and more.

Dataset

• Contains weather-related features and a target variable RainTomorrow (0 = No rain, 1 = Rain).
• Class imbalance was handled by oversampling the minority class.
• Data source: weatherAUS.csv

Models Used

• Logistic Regression
• Random Forest
• Gradient Boosting
• Linear SVM (with calibration)
• K-Nearest Neighbors (KNN)

Models are evaluated using Accuracy and ROC-AUC score.

Key Features

• Missing values were handled by using the mode for categorical variables and the median for numerical variables
• Oversampling to handle class imbalance
• Feature scaling with Min-Max Scaler
• Visualization of correlations and distributions
• Model training, evaluation, and comparison