course-ML

This repository contains a machine learning project for a university course, focused on predicting car insurance claims using a real-world dataset.

Project Overview

The main script, part_B_group_7.py, demonstrates the full workflow of a supervised classification task, including:

• Data Preprocessing: Handling missing values, feature engineering, and categorical encoding for variables such as driving experience, credit score, annual mileage, and more.
• Exploratory Analysis & Visualization: Includes 3D and 2D plots to analyze decision tree parameters and visualize clustering results.
• Modeling: Implements and compares several machine learning models:
  • Decision Tree Classifier (with hyperparameter tuning and visualization)
  • Neural Network (MLPClassifier, with grid search for optimal parameters)
  • XGBoost Classifier (with manual and grid search hyperparameter tuning)
  • KMeans clustering and PCA for unsupervised analysis
• Evaluation: Uses metrics like F1 score, accuracy, and confusion matrices to assess model performance.
• Prediction: Generates final predictions for test data after preprocessing and model selection.

Structure

• part_B_group_7.py: Main code for data processing, modeling, and evaluation.
• data/: Contains training and test datasets.
• grid search results & final predictions/: Stores results from hyperparameter tuning and final model predictions.
• part B group 7.pdf: Project report.

This project showcases practical machine learning techniques for tabular data, including feature engineering, model selection, and performance analysis.