A Novel Approach Based on an Improved Random Forest to Forecasting the Air Quality of Second-Hand Housing

Accurately learned air quality of second-hand housing is very important for home buyer. In order to provide home buyer with a simple and novel approach to get air quality of second-hand housing, this paper firstly presents air quality of second-hand housing forecast method using random forest that is one of the data mining algorithm. According to hedonic price theory, by training the random forest model, this method establishes the internal mapping relationship between features that include prices of second-hand housing and air quality of second-hand housing. Then, based on above model, an improved random forest algorithm is presented to improve the forecast accuracy. Air quality of second-hand housing is predicted exactly by features. At last, empirical studies, using the data set of second-hand housing in Tianhe district, Guangzhou city, show that our proposed method is practical and effective in prediction of air quality of second-hand housing.