| Housing price is an important measure of the economy and greatly impacts an individual’s assets. This paper mainly studies the Shanghai housing price regression in January 2022. The data is based on home listing prices collected by Lianjia, which includes 17,286 individual homes in 4,686 real estate districts. There are 7 variables at the individual house level and 28 at the neighborhood-level, which is important for housing prices. Feature importance is calculated to value its weight in influencing house prices. Random forest, support vector regression, K-nearest neighbor regression, and a self-built binary model will be used. R2, RMSLE, and prediction distribution map will be used to measure the model. This project finds that random forest has the best performance in these models. Then, based on the random forest model, 15 features are selected, and the average housing price trend from 2001 to 2020 is taken as the scale price to build the housing price model and forecast the housing price with an accuracy of over 90%. I will also explain some of the reasons why these factors will influence housing prices in a more macro way. |
