Computer mathematical statistics applied in the housing price investigation through machine learning and linear regression model

Abstract

To Figure out the inner law of the housing price in Beijing and the key factors that influence the housing price, quantitative analysis is used through three statistical tools: autoregressive moving average model, linear regression model and random forest model. The first model shows that the coefficient of autoregressive part is greater than one, revealing an increasing trend of housing price. The second model shows that there are four factors, number of residents in Beijing, production of tertiary industry, per capita urban disposal income, number of elementary school students, have strong linear relationship with the housing price. From the third model, there are three factors, number of subways, number of buses and number of medical institutions, may have strong non-linear relationship with housing price. Then the conclusions that the increasing trend of housing price in Beijing may be kept in the future, there are 7 main factors having robust relationship with the housing price in Beijing, though some of them may not be linear and the statistical tools can be really helpful in analyzing the field of housing price or some other fields have been reached.