Modeling Property Prices Using Neural Network Model for Hong Kong
Start Page / End Page
Xin J. Ge, G. Runeson
121 / 138
International Real Estate Review
This paper develops a forecasting model of residential property prices for Hong Kong using an artificial neural network approach. Quarterly time-series data are applied for testing and the empirical results suggest that property price index, lagged one period, rental index, and the number of agreements for sales and purchases of units are the major determinants of the residential property price performance in Hong Kong. The results also suggest that the neural network methodology has the ability to learn, generalize, and converge time series.