The main goal of this paper is to study and compare the predicting power of Hedonic pricing model and Artificial Neural Network (ANN) model and determining the optimum model for forecasting the housing prices for the city of Tabriz. The results of the estimation of the price function show that most of the variables in the Hedonic pricing model are significant with the expected signs, according to the theory. Physical factors have more important effect than locational factors (environmental and accessibility) on the housing prices. Among the structural characteristics, having lobby, swimming pool, number of bedrooms, and the frontage of the building are the most important factors affecting the prices. The most important locational characteristic on the housing price is the distance to educational centers. For our comparison, we have utilized MSE, RMSE, MAE and criteria. According to all criteria, ANN model had less error in prediction of the housing Hedonic prices for the city of Tabriz. In order to test the hypothesis of equal predicting power of the two models, we have used MGN test. The result of this test indicates that the ANN approach is statistically superior to hedonic model.
Khalili Araghi M, Nobahar E. Predicting Housing Prices for the City of Tabriz: Application of the Hedonic Pricing and Artificial Neural Network Models. qjerp 2012; 19 (60) :113-138 URL: http://qjerp.ir/article-1-189-en.html