Study On The Batch Evaluation Of Commercial Real Estate Price Based On Internet Data

Real estate industry is the basic,leading and pillar industry of national economic development.Real estate is an important part of resident assets in China.Commercial real estate is one of the main forms of real estate.Its price evaluation plays an important role in holding,leasing,transferring,mortgage,requisition,judicial auction,tax collection,enterprise asset restructuring,restructuring,listing and other activities.In fact,the evaluation of commercial real estate has become one of the main contents of our economic and social life.Real estate evaluation can be divided into single evaluation and batch evaluation according to the number of evaluation objects.Single evaluation usually refers to the evaluation of a single subject,while the evaluation of a series of subjects is batch evaluation according to the standards.The current commercial real estate in our country generally uses single evaluation methods,most of which are market comparison method,income method and cost method.Their shortcomings are different standards and low efficiency.Using Internet data and spatial geographic information,introducing machine learning and other methods to implement batch evaluation can effectively improve the traditional evaluation methods.The study of real estate batch evaluation in China mainly focuses on housing,usually using hedonic price theory and machine learning method to achieve batch evaluation.Because commercial real estate has the characteristics of low-frequency transaction and scarcity of samples,the research and application of commercial real estate batch evaluation are relatively insufficient.This paper draws lessons from some methods and paths of housing evaluation,uses the theory of characteristic price to construct a multi-source price fusion model,and carries out an empirical study of commercial real estate batch evaluation.The main work includes:(1)through the study of existing literature,38 characteristic factors affecting the price of commercial real estate are sorted out from four aspects of transportation,commerce,neighborhood and construction,and the corresponding quantitative methods;(2)using machine learning method to quantitatively study these price characteristics and the validity of commercial real estate price prediction;(3)designing a multi-source data provider;Industrial real estate price forecasting model realizes the effective integration and utilization of price data from third-party evaluation agencies and price data obtained on the Internet;(4)Taking 4200 shops in 9 districts of Banan District of Chongqing as the research object,520 samples were selected and processed by data grouping and normalization,using four kinds of regression,including multiple linear regression,exponential linear regression,support vector regression and multi-source price fusion.The batch evaluation experiment was carried out.The experimental results show that the price characteristic variables and multi-source price fusion model designed in this paper can accurately evaluate the batch of specific commercial housing(shops).