Analysis Of Changsha Second-hand Housing Market Based On Data Mining

Since the prevention and control of the epidemic,affected by multiple factors,China’s real estate industry has experienced a downward trend after the adjustment period.However,since 2022,due to the concerted efforts of all parties,the decline has narrowed,and the operation of the real estate market has shown relatively stable and positive changes.It has become a new growth point and consumption hotspot of the regional economy.The real estate industry has distinct regional characteristics.Housing prices are an important factor influencing urban settlement,and second-hand housing has also attracted the inclination and attention of more and more settler.Changsha is a strategic city for economic development in the Changsha-Zhuzhou-Xiangtan delta region.The research on the second-hand housing market is of great significance to the sustainable and stable development of Changsha’s real estate.Based on the crawler technology,this paper obtains the transaction data of secondhand houses for sale in Changsha as the research sample.Firstly,data preprocessing and correlation analysis are carried out on the second-hand housing transaction data that has been climbed,and the correlation between each characteristic variable and house price is explored.At the same time,the statistical analysis chart is used for data visualization analysis,so as to have a preliminary understanding of Changsha’s secondhand housing market.To understanding.Secondly,the Pearson correlation coefficient and the random forest algorithm are used to sort the importance of the feature variables,the embedding method is used to incorporate the feature variables into the model in batches for modeling,and the grid search algorithm is used to optimize the model parameters.Then,this paper uses machine learning random forest,XGBoost regression and Light GBM regression and deep learning LSTM regression model to model the housing transaction price of second-hand houses for sale in Changsha.The results show that the deep learning LSTM algorithm is compared with the three algorithms of machine learning.The error in the price prediction of second-hand housing in Changsha studied in this paper is the smallest,the root mean square logarithm error is 0.2012,and the model fitting effect is the best.This paper uses the transaction data of second-hand houses in Changsha to analyze the impact of various variables on house prices,and further establishes a second-hand house price prediction model to predict the price.The house price forecasting model which has high accuracy.