Forecasting Residential Property Prices in Jiaxing City, China: A Hybrid Machine Learning Framework Integrating Gaussian Process Regressions and Bayesian Optimization

China’s property market underwent a period of rapid growth during the past decade, followed by a significant downturn commencing in late 2021. These systemic shifts, prompted by changing macroeconomic landscapes, have introduced considerable complexities in predicting real estate valuations for regulatory institutions and industry participants. To mitigate these methodological limitations, this research employs Gaussian Process Regression (GPR) incorporating diverse kernel configurations and basis functions to examine monthly fluctuations in residential property values within Jiaxing City, Zhejiang Province, utilizing a comprehensive dataset covering October 2011 to July 2024. The framework was enhanced through Bayesian hyperparameter optimization and cross-validation techniques, demonstrating robust predictive accuracy in out-of-sample evaluations (January 2022–July 2024), evidenced by a Relative Root Mean Square Error (RRMSE) of 0.2523%. The results advance theoretical understanding of urban housing market mechanisms while offering practical implications for policy formulation, whether implemented autonomously or in conjunction with supplementary predictive approaches.