Real Estate Segmentation: A Model of Real estate Decision Support System

Abstract

Due to human limitations of computational thinking, the quality of rational decision-making is constrained, and as a result, people encounter bounded rationality. A decision support system is widely demanding in tackling this problem, especially in real estate management. This study focuses on 3 main purposes. Firstly, comparing K-means, X-means and K-medoid algorithm performance in clustering sold house characteristics to be further used for pricing houses. Second, characterizing each cluster for developing a suitable marketing strategy by utilizing machine learning technology. Lastly, providing a managerial implication as a decision support system for assisting stakeholders in making a decision. Eventually, K-means and X-means algorithm show very similar performance. X-means can automatically determine the number of clusters while k-means utilize the elbow method to find the optimum number of clusters. Three clusters were identified as cluster 0, cluster 1, and cluster 2. Cluster 0 was occupied by 85.77% of low house prices. There are two practical implications of this study. Firstly, the results of clustering analysis which reflected in a model of decision support system. Second, an intuitive and comprehensive methodological framework is presented for helping stakeholders designing a decision support system.