Comparison of data driven models for the valuation of residential premises using KEEL

  • Authors:
  • Tadeusz Lasota;Jacek Mazurkiewicz;Bogdan Trawiński;Krzysztof Trawiński

  • Affiliations:
  • Department of Spatial Management, Wrocław University of Environmental and Life Sciences, ul. Norwida 25/27, 50-375 Wrocław, Poland;Institute of Computer Engineering, Control and Robotics, Wrocław University of Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland;(Correspd. E-mail: bogdan.trawinski@pwr.wroc.pl) Institute of Informatics, Wrocław University of Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland;European Centre for Soft Computing, Edificio Científico-Tecnológico, 3$^a$ Planta, C. Gonzalo Gutiérrez Quirós S/N, 33600 Mieres, Asturias, Spain

  • Venue:
  • International Journal of Hybrid Intelligent Systems - Hybrid Fuzzy Models
  • Year:
  • 2010

Quantified Score

Hi-index 0.01

Visualization

Abstract

The experiments aimed to compare data driven models for the valuation of residential premises were conducted using KEEL (Knowledge Extraction based on Evolutionary Learning) system. Twelve different regression algorithms were applied to an actual data set derived from the cadastral system and the registry of real estate transactions. The 10-fold cross validation and statistical tests were applied. The lowest values of MSE provided models constructed and optimized by means of support vector machine, artificial neural network, decision trees for regression and quadratic regression, however differences between them were not statistically significant. Worse performance revealed algorithms employing evolutionary fuzzy rule learning. The experiments confirmed the usefulness of KEEL as a powerful tool with its numerous evolutionary algorithms together with classical learning approaches to carry out laborious investigation on a practical problem in a relatively short time.