DBPapers
DOI: 10.5593/SGEM2014/B22/S9.061

TAXONOMY OF REAL ESTATE PROPERTIES WITH THE USE OF K-MEANS METHOD

B. Calka, E. Bielecka
Wednesday 1 October 2014 by Libadmin2014

References: 14th International Multidisciplinary Scientific GeoConference SGEM 2014, www.sgem.org, SGEM2014 Conference Proceedings, ISBN 978-619-7105-11-7 / ISSN 1314-2704, June 19-25, 2014, Book 2, Vol. 2, 489-496 pp

ABSTRACT
The article presents taxonomy of real estate properties, namely premises, with use of kmeans method. The developed approach separates groups (clusters) of premises with similar characteristics, which do not depend on spatial location. The research was carried out for Siedlce, the city located in the middle east of Poland. The analysis involved 1873 residential properties which were subjects of transaction in 2007-2011. In grouping procedure, the following characteristics were taken into account: floor, area, date of construction, distance from the town centre and standard of the premises. The necessity of determination a priori the number of clusters forced to use agglomeration method before clustering. The distinguished 7 clusters group premises with similar non spatial characteristics affecting the value of the premises. However, they differ in price which varies between 3295-3979 PLN for 1 m2. Homogeneity of premises in each group allows elaborate tax map using geostatistical interpolation. The obtained results could influence considerably the effectiveness of properties resources management as well as facilitate decision making.

Keywords: cluster analysis, agglomeration method, k-means method, real estate valuation