Integrated Clustering Regression for Real Estate Valuation

This paper presents an integrated clustering and regression approach for real estate valuation.The location of a property represents an essential factor that is usually included in real estateappraisal. Considering the spatial distribution of the proprieties in the dataset under study withrelative unit price, it is motivated to apply the clustering method to group similar proprietiesin homogeneous groups. Then, conduct regression analysis for prediction within each cluster.The used approach shows that predictions improved slightly for one cluster and worsens forthe other cluster. The possible reasons for such performance and suggested strategies to testthe methodology were discussed