Research on Big Data Mining and Analysis of Urban Housing Real Estate Based on Machine Learning
This paper points out that the mining and analysis of urban housing real estate data is very conducive to the analysis of the trading rules of urban housing real estate,but the advantages and disadvantages of various prediction algorithms are not clear.Therefore,through the big data analysis of the transaction price of second-hand housing in a city,the prediction effects of four second-hand housing prediction algorithms are compared.The results show that when the housing area is less than 90 square meters,the transaction price is mostly less than the average price;when the housing area is between 90 and 120 square meters,the transaction price of the house is basically maintained near the average price;when the housing area is greater than 120 square meters,the transaction price of the house is very discrete.The average absolute error,mean square error and root mean square error of the four prediction algorithms from large to small are:Lasso model>Random Forest Regressor algorithm>XGBoost algorithm>Stacking algorithm;the average absolute error,mean square error and root mean square error of the Stacking prediction algorithm are the smallest among the four algorithms,and the accuracy is the highest.Therefore,the Stacking prediction algorithm is more suitable for predicting the price of second-hand housing.The research results can provide reference for the decision-making of urban housing real estate transactions.
万方数据
维普
