Research On Prediction Of Commodity Housing Price Index Based On Web Search Keywords

With the development and popularity of the Internet,people's consumption habits have changed.People are accustomed to querying relevant information through search engines before making decisions,especially consumer decisions.Therefore,the search keywords represent people's attention hotspots to a certain extent,and the search volume can represent the degree of attention.It can be concluded that the network search keyword data contains the interest of the market participants,representing the trend of its behavior,and can provide the necessary microscopic data for solving macroeconomic problems.As the real estate industry is a key industry to promote the rapid development of the national economy,more and more research scholars pay attention to real estate prices.This paper takes Kunming as an example to study the sales price index of new commercial housing and network search keyword data.Based on the real estate market supply and demand theory,expectation theory and behavior theory,this paper first analyzes the main influencing factors of residential prices,and analyzes the relationship between the sales price of new commercial housing and the network search keyword data from a qualitative perspective.According to the influencing factors of housing price,using the core keywords and expanding the keyword selection method,the network search keywords are screened,and a thesaurus containing 45 keywords is formed initially;then the gray correlation analysis method is used to quantitatively the relationship between the sales price index of new commercial residential buildings in Kunming and the network search keyword data is analyzed.Next,using Spearman correlation analysis to screen out high-correlation search keywords,and using the principal component analysis method to synthesize keyword indicators,form eight comprehensive indicators.Finally,a neural network prediction model is established based on the comprehensive index as the data input.The research conclusions mainly include the following three points:(1)There is a strong correlation between the network search keywords and the price of new commercial housing in Kunming;(2)The forecast model established by the search keyword data can predict the sales of new commercial housing in Kunming.Price index;(3)Using the network search keyword data to predict the residential price index has a good timeliness.