A Multi-Attribute Sales Comparison Method for Real Estate Valuation

Abstract

The theory and practice of real-estate valuation has attracted immense interest over the past decades. This paper is concerned with the sales comparison approach. First, a brief survey of some procedures used worldwide, the sales comparison by adjustments, the hedonic regression and the hedonic price index method, is presented. To improve the versatility of property appraisals a new valuation method, as a combination of multi-objective optimization (MOO) and multi-criteria decision analysis (MCDA) is developed. A unique feature of this model is that it enables the appraiser to evaluate the characteristics of a property on those scales of measurement to which they belong de facto. To comply with this objective, distinct metric distance functions on each scale of measurement, including the two qualitative scales (nominal and ordinal), are employed. After that, the physical worth of a property is derived as a weighted sum of the composite scores. This can be measured on an interval scale. To predict the monetary worth of a property, a simple linear regression model is developed. The benefits of the use of this multi-attribute valuation method is also discussed. A comprehensive real-world study showing the application of the procedure is included.