Forecasting Hong Kong Property Prices: Multiple Regression Method vs An Artificial Neural Network Approach

World of Construction Project Management
Publication Type:
Conference Proceeding
Proceedings of the 1st International Conference, World of Construction Project Management, 2004, NA pp. 79 - 86
Issue Date:
Filename Description Size
Thumbnail2006009426OK.pdf1.47 MB
Adobe PDF
Full metadata record
Hong Kong's residential property market has experienced significant changes over the last two decades. It is an interesting challenge for modeling the volatility of property prices. The aim of the study is to build forecasting models for forecasting the private residential property prices using macro data in Hong Kong. Both artificial neural network approach and multiple regression analysis are employed for comparing the forecasting results. Variables such as lagged property prices, household income and transaction volume are derived to test the models. The results show that both methods are valid and that the artificial neural network demonstrates a good prediction power with low forecasting error.
Please use this identifier to cite or link to this item: