The Appraisal Research Of Office Buildings Based On Random Forest

With the development of economy,the relevant economic activities of the tertiary industry in China are more and more frequent,and the demand and supply of office buildings are increasing year by year.As there is less and less land available for development in the city,the trading of office buildings becomes more active,and the demand for the evaluation of office buildings also increases accordingly.Whether from the perspective of market participants,or from the point of view of the state levy taxes,the precise measurement of office prices is an eternal topic.At present,the three traditional evaluation methods,market comparison method,cost method and income method,rely too much on the experience of the evaluator in the actual application,and seldom use mathematical models,so the cost is high in the actual application.Drawing on the experience of foreign real estate appraisal,this paper introduces the theory of hedonic price to understand the housing price from the perspective of consumer demand for housing characteristics,and conducts the housing price evaluation research.In the process of constructing office building evaluation model by using hedonic price theory,multiple linear regression method is often used for regression prediction.The selection of function form has a great influence on the evaluation effect,but it often relies on the artificial linear assumption,which is easy to cause a large error.This paper introduces the latest research result of machine learning theory--random forest method to make regression prediction of hedonic price model and establish the price evaluation model of office building.Finally,the office appraisal model is applied to the empirical research along the urban rail transit in Chongqing.On the basis of the office listing data,the hedonic price data of 210 office buildings in the study area are obtained through investigation and interview,and then the modeling and prediction are carried out.Finally,the paper uses the commonly used multiple linear regression method and compares it with the prediction model established by the random forest regression method.It shows that the prediction results of the office building appraisl model based on the random forest method have high accuracy and are worth popularizing and applying.