Research About Housing Demand Prediction In Third-Tier City

At present,the housing market of Third-tier city has begun to take shape.Because of the strong correlation between housing industry and other industries,the healthy development of housing market in third-tier city has large impact on regional economic development.But the research about housing demand prediction of third-tier city is less,which will lead the government,developers and residents hard to grasp the housing market objectively.The article takes the eastern Third-tier city of Xuzhou as an example,analyze the factors of housing demand and forecast the housing demand systematically,which can help the government departments to grasp the present situation of housing market and make the land supply more scientific and reasonable.Besides,it can help the real estate enterprises to develop the project planning and save capital.At last,it can guide residents to make the housing consumption reasonably.The article takes the study of influencing factors about First-tier cities and secondtier cities as the reference,and selects 11 indicators of influencing factors for constructing indicator system of influence factors about housing demand in Xuzhou City.After studying the applicability of each indicator in Third-tier City,the paper improves the original indicator system,and finally constructs the new indicator system of Third-tier City.After that,the paper uses the GM(1,1)model,multiple stepwise regression analysis and BP neural network model to forecast the housing demand in Xuzhou city.After analyzing the applicability,accuracy of the prediction results of each model,the article confirm the predicting value of housing demands in Xuzhou city from 2016 to 2020 finally.The results prove that:(1)Urban per capita disposable income,urban per capita consumption expenditure,average wages of staff and workers,urban population,urbanization rate,GDP and residential investment have strong correlation with housing demand(correlation>0.7),they belong to the key influencing factors;(2)Indictors of urban population and urbanization rate can reflect the impact of regional population size and urbanization level on housing demand in Third-tier city,which is better than regional population and the sale quantity of commercial buildings;(3)Of these three kinds of forecasting methods,the prediction precision of multiple stepwise regression analysis model is the highest(the average relative error is 5.69%),BP neural network model in the second place(the average relative error is 8.49%),and the GM(1,1)model in the end(the average relative error is 10.76%);(4)The predicting process of GM(1,1)model does not consider the influence of related factors,but the multiple stepwise regression analysis and BP neural network take the key factors as independent variables to build prediction model,which predicting process and results become more scientific and reasonable;(5)Due to the smaller average relative error and the similar trend characteristics of multiple stepwise regression analysis and BP neural network,the article finally takes the prediction results of multiple stepwise regression analysis and BP neural network as the upper and lower limits of housing demand in Xuzhou City from 2016 to 2020,which can reduce probable deviation of using single method;(6)There will be a stable growth of the future housing demand in Xuzhou City,and the development prospect of housing market will be more optimistic,but there are still some issues to resolve.The article puts forward some suggestions such as strengthening of economic construction,strengthening the data collation work,improving the data disclosure mechanism,adjusting the policy guidelines,improving the regulation of housing market and other aspects;(7)The conclusions of this paper can provide reference for the local government,the real estate developers and citizens,which can also provide reference for other third-tier cities in China.