A REVIEW ON THE PERFORMANCE OF HOUSE PRICE INDEX MODELS: HEDONIC PRICING MODEL VS ARTIFICIAL NEURAL NETWORK MODEL

Abstract

Most of the countries around the globe including Malaysia using the Hedonic Pricing Model (HPM) to build a national house price index. However, HPM seems to not reflect the current market trend due to its non-linearity and multicollinearity. Hence, one of the alternative techniques that have been applied is an artificial neural network to improve the reliability of forecasting information. The present article conducted a systematic literature review on the artificial neural network in forecasting local house price index. This study integrated multiple research designs and the review was based on the publication standard, namely ROSES (RepOrting Standard for Evidence Syntheses). This study selected articles through various databases such as Scopus, Emerald, Science Direct, SpringerLink, Web of Science, and Google Scholar. This review has five main themes namely 1) The Malaysian House Price Index (MHPI); 2) Hedonic Pricing Model; 3) Main problem of hedonic price function; 4) Artificial Neural Network (ANN), and 5) Performance of ANN. The findings explained the performance of artificial neural network in forecasting house price index based on previous study. This study benefits to the expansion of academic knowledge on property valuation in Malaysia where this research explores for improvement in the valuation sector using a reliable approach that is ANN model.