Analysis Of Influencing Factors Of Urban Rental Price Based On Machine Learning Methods

Under the actual conditions of high housing prices in China,renting a house has become an effective way to solve the housing problem,which has always been the focus of livelihood issues.The influencing factors of rental housing prices and reasonable rental prediction models are of positive significance to the healthy and stable development of the rental housing market.On the one hand,it can provide an effective reference for people who need rent a house and help them choose more cost-effective housing;on the other hand,it will help landlords and housing agencies make more reasonable rent pricing.We select a total of 10,672 housing rental data in Beijing,Shanghai,Hangzhou and Tianjin from a housing rental and sales website.First,various statistical charts are used to discuss the influence of building features,location features,and environmental features on housing rent after feature selection.Secondly,we establish four city rent prediction models based on random forest and artificial neural network.We also rank the importance of the features.It can be seen from the ranking results that in terms of architectural characteristics,the area of the house is the factor that has the greatest impact on rent.Where the house is located and the distance from the nearest subway station have a great impact on the rent in terms of location characteristics;in terms of environmental characteristics,the type of the house also has a significant impact on rents.The conclusion is drawn by comparing the prediction results of the two models that the prediction effect of the random forest model is better.Finally,we use the overall data of the four cities to construct an XGBoost model and obtain reasonable rent prediction results.