Abstract
Research background: The value of the property can be determined on an individual or mass basis. There are a number of situations in which uniform and relatively fast results obtained by means of mass valuation undoubtedly outweigh the advantages of the individual approach. In literature and practice there are a number of different types of models of mass valuation of real estate. For some of them it is postulated or required to group the valued properties into homogeneous subset due to various criteria. One such model is Szczecin Algorithm of Real Estate Mass Appraisal (SAREMA). When using this algorithm, the area to be valued should be divided into the so-called location attractiveness areas (LAZ). Such division can be made in many ways. Regardless of the method of clustering, its result should be assessed, depending on the degree of implementation of the adopted criterion of division quality. A better division of real estate will translate into more accurate valuation results.
 Purpose of the article: The aim of the article is to present an application of hierarchical clustering with a spatial constraints algorithm for the creation of LAZ. This method requires the specification of spatial weight matrix to carry out the clustering process. Due to the fact that such a matrix can be specified in a number of ways, the impact of the proposed types of matrices on the clustering process will be described. A modified measure of information entropy will be used to assess the clustering results.
 Methods: The article utilises the algorithm of agglomerative clustering, which takes into account spatial constraints, which is particularly important in the context of real estate valuation. Homogeneity of clusters will be determined with the means of information entropy.
 Findings & Value added: The main achievements of the study will be to assess whether the type of the distance matrix has a significant impact on the clustering of properties under valuation.
Highlights
There are two main approaches in real estate valuation: individual valuation and mass valuation
The plots of land subject to analysis were qualified into location attractiveness zones (LAZ) by means of agglomeration clustering with spatial constraints
The study covered over 1600 plots of land, which were subject to mass valuation with the use of the Szczecin Algorithm of Mass Property Valuation
Summary
There are two main approaches in real estate valuation: individual valuation and mass valuation. In the case of mass valuation, the subject of valuation is a large number of properties of one type, which are appraised with a uniform approach yielding consistent results. One of the basic elements of many mass valuation models is the division of the valued area into subareas, which in the case of valuations for tax purposes are called tax zones. The specification of these zones constitutes one of the key problems from the point of view of the correctness of the obtained valuation results. If there are properties in a given zone exerting different influence of features on the value, the obtained valuation results will be inaccurate. The issue of proper specification of valuation zones is an important economic and computational matter
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
