This repository contains a Flask web application that uses a CatBoost regression model to predict earthquake magnitudes. The model is trained using a dataset containing various features related to earthquakes. This project demonstrates the process of loading data, preprocessing it, training a machine learning model, and deploying the model as a web service using Flask.
- Data Preprocessing: Load and clean the dataset, handle missing values, and prepare categorical features.
- Model Training: Train a CatBoost regression model using k-fold cross-validation for robust performance evaluation.
- Model Evaluation: Evaluate the model using metrics such as Mean Squared Error (MSE), R-squared (R²), and Mean Absolute Error (MAE).
- Web Interface: Provide a user-friendly web interface to input features and get earthquake magnitude predictions.
- Model Persistence: Save and load the trained model using pickle for easy reuse.
app.py: The main Flask application file containing routes and model handling code.dataset.csv: The dataset file (not included, needs to be added by the user).catboost_model.pkl: The serialized CatBoost model file (generated after training).templates/index.html: The HTML template for the web interface.requirements.txt: The list of Python dependencies required to run the application.
- Flask
- numpy
- pandas
- scikit-learn
- catboost
- pickle
Feel free to contribute, provide feedback, or report issues related to this project.