The Machine Learning Based Regression Models Analysis for House Price Prediction

Abstract

House Price Prediction focuses on the development of methods that use ML algorithms to accurately predict house prices. Machine learning has been expanding at a faster rate this decade. There are a lot of new algorithms and uses for Machine Learning every day. Among these applications, one may find journal articles on home price prediction. The need to model home price prediction has arisen due to the fact that property prices are growing annually. Customers may utilise these built models to choose the perfect home for their needs. This study explores the feasibility in predicting house prices concerning the 2019 data obtained from the property listings in Kuala Lumpur by using machine learning algorithms on the gathered data set. Data cleaning, data transformation, and data feature selection were performed sequentially on the dataset to get a clean dataset ready for training the models. These include RF, SVM, LightGBM, and GBR models to assess their prediction model. Hence, the Gradient Boosting Regressor model proved to be the best model since it produced the lowest MSE, RMSE, and the highest R2 score compared to the other models. The experimental results show the GBR model gets a higher 90% accuracy for house price prediction in terms of other models. The findings of this study can be used to inform stakeholders in the real estate, urban planning and investment markets regarding the need for high-quality data combined with suitable algorithms to assist in real estate price prediction.