基于随机森林模型的重庆市二手房价格预测研究

Research on Price Prediction of Second-Hand Housing in Chongqing Based on Random Forest Model

摘要: 二手房价格的精准预测对购房者和政府对房地产政策的调控具有重要意义，本文以重庆市九大主城区2015~2020成交的二手房价格为依据，将影响房价的微观因素与宏观因素有机结合，运用Python使用随机森林算法对二手房数据集进行训练建模，最后结合岭回归、Lasso回归对训练结果进行比较，实验结果显示随机森林模型的误差最小，应用效果比较好，值得推广和应用到房地产价格评估中。

Abstract: The accurate prediction of the second-hand house price is of great significance to the buyers and the government’s regulation of the real estate policy. This paper takes the prices of second-hand houses that were traded between 2015 and 2020 in the nine major urban areas of Chongqing as reference, organically combines the micro and macro factors that affect the housing prices, and uses Python and random forest algorithm to conduct training modeling on the second-hand house data set. Finally, ridge regression and Lasso regression are combined to compare the training results. The experimental results show that the error of random forest model is the smallest, the application effect is better, and it is worth popularizing and applying to the real estate price evaluation.