Author(s)

N. Y. Zainun, I. A. Rahman & M. Eftekhari

Abstract

Over the past decade, the rate of growth of housing construction in Malaysia has been dramatic. The level of the urbanization process in the various states in Peninsular Malaysia is considered to be important in planning for low-cost housing needs. The aim of this study is to develop a Neural Networks model to forecast low-cost housing demand in Batu Pahat, Johor, one of the states in Peninsular Malaysia. The time series data was analyzed using Principal Component Analysis to determine the significant indicators which will be the input in Neural Networks model. The feed forward network with the most commonly used training algorithm, back propagation networks is used to develop the model. The results show that the best Neural Network model is 2-25- 1 with 0.7 learning rate and 0.4 momentum rate. Neural Networks can forecast low-cost housing demand in Batu Pahat very well with 0% of MAPE value. Keywords: forecasting, low-cost housing, artificial neural networks. 1 Introduction In each five year National Plan, Malaysia’s government has focused on various housing programmes to ensure that all Malaysians, particularly the low income groups, have access to adequate and affordable shelter and related facilities [1]. During the Ninth Plan period, the development of the housing sector continues to focus on the provision of adequate, affordable and quality houses for all Malaysians [2]. The housing is divided into four main categories; low cost, low medium cost, medium cost and high cost housing. In Malaysia low cost

Keywords

forecasting, low-cost housing, artificial neural networks