Research On Housing Rent Prediction Based On Data Mining

The price of commercial residential building is continuing to rise now.The high price of housing causes the rental life model to be increasingly favored by people.There is the information asymmetry between landlords and tenants in the rental market,resulting in the waste of housing resources to some degree.By the machine-learning algorithms,this paper establishes a housing rent prediction model to forecast the rental price,holds the forecast analysis to the housing rent price and provides reference for the landlord's reasonably pricing the rent and the tenant's obtaining the cost-effective housing.The detailed work of this paper is as follows:1.Data acquisition and preprocessing.the online rental housing data of Guiyang City(Yunyan District,Baiyun District,Huaxi District and other four districts)was collected from Lianjia.The original data set was handled by data cleaning and feature transformation.Hence,data can be effectively used for the established model.2.Descriptive statistics.The histogram is used to describe the distribution of housing rent,and the box chart is used to view the influence of housing construction and surrounding traffic conditions on housing rent in the administrative area.3.The random forest was used to sort the importance of each feature of the sample data and the importance of each characteristic variable on the rent was analyzed.While the model was trained,the Wrapper method was combined with the specific model to select the characteristic subset suitable for the specific model as the final input feature set for modeling.4.Building the housing rent forecast model.Use Random Forest,Multiple Linear Regression,XGBoost,LightGBM and Deep Neural Network on the basis of the feature selection to forecast the housing rent and use the Root Mean Square Error(RMSE)as a model assessment criteria and respectively determine the model in the performance of the sample data sets,Through comparing the built five regression models,LightGBM model was the optimal in the data set of housing rent.This paper,by using online rental housing data,analyze the index on the fluctuation of house rent,and further rent prediction model is established to forecast the online house rent.Experiments show that the prediction model established in this paper can better forecast housing rents.