A House Price Prediction Model Based on K-means Clustering and Random Forest in Guangzhou

Abstract

This paper addresses the key issues in house price forecasting from multiple perspectives by establishing a house price forecasting model for Guangzhou city, providing valuable information and decision support for home buyers, developers, and the government. First, this paper employs the Person coefficient, stepwise regression model and t-test to address the problem of quantifying data and exploring house price factors. By analyzing the correlation between the relevant variables and house prices, the key characteristics that have significant and strong correlation effects on house prices are obtained. Second, the K-means clustering model was used to classify houses into three categories: economic houses, comfortable houses and high-end houses. This classification result provides a more detailed data base for the subsequent construction of the house price prediction model. Finally, the random forest house price prediction model was established in this paper, and the model was validated by error analysis and stability analysis. The average absolute value error and goodness-of-fit obtained were 0.08 and 0.92, respectively, indicating that the model has high accuracy and reliability. The research in this paper has important implications for all parties, including home buyers, developers, and the government. For home buyers, the model can help them better understand the market situation; for developers, the model can guide their reasonable pricing and development strategies; for the government, the model can provide a scientific basis for real estate market regulation and policy control to promote market stability and sustainable development.