| Predicting the market trend is an important step to maintain market stability.Therefore,the mathematical model is applied to the market to quantify market changes with data to facilitate a clearer understanding of market changes,which will help improve the accuracy of market research.This paper mainly builds a model of real estate market based on multiple regression model and VAR model.Because the real estate market has the dual attributes of consumption and investment,the market changes are closely related to many factors,so this paper chooses the main influencing factors for variance contribution analysis,and preliminarily determines the influence ability of each factor.Then,in order to grasp the market changes more carefully,the population factors are divided into different levels,and the model is analyzed again.In view of the natural and social attributes of population factors,this paper chooses five explanatory variables: sex ratio,dependency ratio,the proportion of non-illiterate people,the proportion of families with less than three people and the proportion of spouses.The model is established by using the quarterly data from 2002 to 2017.In the process of modeling,ADF stationarity test and cointegration test are used to confirm that the data sequence can be applied without the problem of pseudo-regression.Considering that the influencing factors in the market are generally interrelated and multi-collinearity,this paper will also establish the regression model by principal component analysis method,and choose the model with better fitting effect.Then,establishing the VAR model.Under the condition of guaranteeing the stability of the model,the time series of variables and the response of the interpreted variables are determined by Granger causality test,impulse response function and variance decomposition.Finally,the market phenomena are explained according to the results of the model to achieve the analysis effect of forecasting market changes. |
PDF Full Text Request