基于机器学习的二手房价格预测研究OA

Second-hand housing transaction price evaluation has important reference value and provides a basis for the decision-making of governments,house buyers,sellers and real estate agencies.This paper takes second-hand house prices in Taiyuan city,Shanxi Province as the research object,constructs prediction models and selects the optimal scheme by comparing model effects.Firstly,Web crawler technology is used to obtain second-hand housing data of Taiyuan from Lianjia website,and 8 394 data records including 63 most representative characteristic variables are finally obtained for house price prediction.Secondly,the matplotlib library of Python is used to conduct visual research from three aspects:location characteristics,building characteristics and transaction characteristics,and the influence relationship between house prices and various variables is initially determined.Finally,in order to select the optimal model to predict the changes of second-hand house prices in Taiyuan,the CART Decision Tree model and the XGBoost model are constructed respectively.The model comparison results show that the XGBoost model has high accuracy and is more suitable for second-hand house price prediction.