brianotiodhiambo-source/analystlab-week6-feature-engineering-model-optimization

Week 6 AnalystLab Africa Data Science Internship Project focusing on Feature Engineering, Feature Selection, Model Optimization, and Hyperparameter Tuning for House Price Prediction using Gradient Boosting Regressor. # Week 6: Feature Engineering & Model Optimization ## AnalystLab Africa Data Science Internship ### Project Overview This project focuses on improving machine learning performance through feature engineering and model optimization using the Housing Dataset. ## Objectives - Create meaningful features from raw data. - Transform variables for machine learning. - Select the most informative features. - Optimize the regression model using GridSearchCV. - Evaluate performance before and after optimization. ## Feature Engineering The following features were created: - Area per Bedroom - Total Rooms - Parking Availability - Luxury House Indicator ## Feature Selection Recursive Feature Elimination (RFE) selected the ten most informative features for model training. ## Model Used Gradient Boosting Regressor ## Hyperparameter Tuning GridSearchCV Best Parameters: - Learning Rate = 0.05 - Max Depth = 2 - Number of Estimators = 100 ## Performance Baseline Model - MAE: 1,047,683.40 - RMSE: 1,401,484.80 - R² Score: 0.6114 Optimized Model - MAE: 1,074,514.60 - RMSE: 1,446,907.33 - R² Score: 0.5858 ## Technologies - Python - Pandas - NumPy - Scikit-learn - Matplotlib - Jupyter Notebook ## Author Brian Otieno Odhiambo