1 Introduction
Housing price forecasting plays a vital role in macroeconomic and financial decision supporting. In the past decade, a global financial crisis has been witnessed, due to inaccurate housing price forecasting and unconscionable financial policymaking [1]. In terms of spatial granularities of predictions, existing studies on housing price forecasting models are mostly on city levels, for supporting macroeconomic analysis and policymaking. The city-level forecasting, however, can not capture the fact of imbalanced development between a city’s mile-level subregions. For instance, in Xi’an, the average real estate prices of three districts in Sept. 2018 increased more than 10 percent while the average prices of the other six districts decreased about 5 percent during the same period [2].