Residential housing price index forecasting via neural networks

During the past decade, the housing market in China has witnessed rapid growth and the significance of forecasting related to housing prices has undoubtedly elevated, which has become an important issue to the people in investment and policymakers in regulations. In this study, we explore neural networks for residential housing price index forecasts from ten major Chinese cities for July 2005-April 2021. We aim at constructing simple and accurate neural networks as a contribution to pure technical forecasts of the Chinese residential housing market. To facilitate the analysis, we investigate different model settings across algorithms, delays, hidden neurons, and data spitting ratios, and arrive at a simple neural network with three delays and three hidden neurons, which produces stable performance of about 0.75% average relative root mean square error across the ten cities for the training, validation, and testing phases. Our results can be used on a standalone basis or combined with fundamental forecasts to form perspectives of residential housing price trends and carry out policy analysis.