Improving the Accuracy of House Price Prediction using Catboost Regression with Random Search Hyperparameter Tuning: A Comparative Analysis

Faezal Hartono (Universitas Dian Nuswantoro)
Muljono Muljono (Universitas Dian Nuswantoro)
Ahmad Fanani (Universitas Dian Nuswantoro)

Advance Sustainable Science, Engineering and Technology (ASSET)
Vol. 6 No. 3 (2024): May - July

Achieving a significant improvement over traditional models, this study presents a novel approach to house price prediction through the integration of Catboost Regression and Random Search Hyperparameter Tuning. By applying these advanced machine learning techniques to the King County Dataset, we conducted a thorough regression analysis and predictive modeling that resulted in a marked increase in accuracy. The baseline model, a conventional linear regression, provided a foundation for comparison, evaluating performance metrics such as R-squared and Mean Squared Error (MSE). The meticulous hyperparameter tuning of the Catboost model yielded a remarkable improvement in predictive accuracy, demonstrating the efficacy of sophisticated data science techniques in real estate and property valuation. The percentage increase in accuracy over the baseline model is explicitly stated in the abstract.