AI/ML Task 2 — Feature Engineering, Model Optimization & Performance Comparison

Internship project at Maincrafts Technology.

What this project does

Builds an enhanced house price prediction system using the California Housing Dataset. Trains three regression models, compares their performance, and selects the best one.

Steps followed

• Loaded the California Housing Dataset
• Applied feature scaling using StandardScaler
• Split data into 80% training and 20% testing
• Trained three models — Linear Regression, Ridge Regression, Decision Tree
• Evaluated using RMSE and R2 Score
• Visualized results with actual vs predicted plot, bar chart, feature importance and residual plot
• Saved the best model using joblib

Results

Model	RMSE	R2 Score
Linear Regression	0.7456	0.5758
Ridge Regression	0.7456	0.5758
Decision Tree	0.7242	0.5997

Best Model: Decision Tree Regressor (max_depth=5)

Files

• AI_ML_Task2_Model_Comparison.ipynb — main notebook
• AI_ML_Task2_Report.docx — methodology and results report
• best_model.pkl — saved Decision Tree model
• scaler.pkl — saved StandardScaler

Libraries used

pandas, numpy, matplotlib, seaborn, scikit-learn, joblib