Estimating Residential Property Values on the Basis of Clustering and Geostatistics

The article presents a two-stage model for estimating the value of residential property. The research is based on the application of a sequence of known methods in the process of developing property value maps. The market is divided into local submarkets using data mining, and, in particular, data clustering. This process takes into account only a property’s non-spatial (structural) attributes. This is the first stage of the model, which isolates local property markets where properties have similar structural attributes. To estimate the impact of the spatial factor (location) on property value, the second stage involves performing an interpolation for each cluster separately using ordinary kriging. In this stage, the model is based on Tobler’s first law of geography. The model results in property value maps, drawn up separately for each of the clusters. Experimental research carried out using the example of Siedlce, a city in eastern Poland, proves that the estimation error for a property’s value using the proposed method, evaluated using the mean absolute percentage error, does not exceed 10%. The model that has been developed is universal and can be used to estimate the value of land, property, and buildings.