Forecasting on House Price Index using Artificial Neural Network

• Main Authors: Gobalkrishnan, Gangaieisvari, Md Yusof, Zahayu
• Format: Article
• Language: English
• Published: Universiti Pendidikan Sultan Idris 2023
• Subjects: QA Mathematics
• Online Access: https://repo.uum.edu.my/id/eprint/30937/1/JCIT%2013%2002%202023%2044-60.pdf; https://doi.org/10.37134/jcit.vol13.2.5.2023; https://repo.uum.edu.my/id/eprint/30937/; https://ojs.upsi.edu.my/index.php/JCIT/article/view/7866; https://doi.org/10.37134/jcit.vol13.2.5.2023
• Institution: Universiti Utara Malaysia

Forecasting the residential property sector is a crucial component in the decision-making process for investors and government in supporting asset allocation, developing property finance plans and implementing a relevant policy. The purpose of this study is to examine the determinants of Penang house price index and to develop a model to forecast Penang house price index in Malaysia. Estimation is done by using ordinary least square and artificial neural network method. Relevant data sets were obtained from the Monthly Statistical Bulletin, Bank Negara Malaysia and National Property Information Centre. The empirical analysis of this research is based on quarterly time series data which cover the periods from 2005Q1 to 2022Q1. The main findings reported that base lending rate and unemployment rate are negatively associated with and have significant impacts on Penang house price index. Meanwhile, gross domestic product is positively related to and has a significant impact on Penang house price index. Consumer price index shows a positive sign; however, it recorded an insignificant impact on Penang house price index. Even though there are three independent variables recorded significant impact on Penang house price index, yet gross domestic product is the most vital determinant of Penang house price index in Malaysia. The artificial neural network model was trained and tested using quarterly time series data from 2005Q1 to 2022Q1 and the model was validated using data from 2021Q1 to 2022Q1. Model validation indicates that artificial neural network has a high level of accuracy in its ability to learn, generalize, and converge time series data efficiently as well as able to generate reliable forecasting information