A Comparative Study of Regression Models for Housing Price Prediction

Abstract

The property market is closely related to the regional economy. This study focuses on exploring the problem of house price prediction in the property market. This paper aims to reveal the performance of extreme gradient boosted tree regression, ridge regression, decision tree, and random forest regression models in feature selection and data processing, and assess their effectiveness in house price prediction. Aspects compared include prediction accuracy, data required for fitting, and running time. The study shows that the random forest and decision tree regression models perform best. Although these two models require a certain amount of data, they are easily satisfied in the context of big data. Although the fit of the random forest regression model and the decision tree regression model is similar, the training time required for the random forest regression model is too long. Therefore, it is concluded that the decision tree regression model performs well on general real estate datasets and is a widely applicable regression model in the real estate domain. This study can provide a helpful reference for data analysis and forecasting in the real estate field and provide a basis and inspiration for further research in related fields.