Random Forest Model for the House Price Forecasting

Abstract

Accurately predicting housing prices is of great significance to improving housing equity. In this work, we first analyzed the factors that affect housing prices. The experimental results show that income is the most important influencing factor, and its correlation with housing prices is as high as 0.69. In addition, room distribution and geographical location are the remaining important image factors. In order to achieve accurate housing price prediction, we designed a random forest to achieve housing price prediction. Compared with support vector regression (SVR), linear regression (LR) and decision tree (DT), our method achieves the best root mean square error (RMSE) and goodness of fit. In contrast, the predicted values of LR and SVR tend to be higher than the true values, while the values of DT tend to be lower than the true values.