A clustering-based approach to land valuation in land consolidation projects


Ertunc E., Karkinli A. E., Bozdag A.

LAND USE POLICY, vol.111, 2021 (SSCI) identifier identifier

  • Publication Type: Article / Article
  • Volume: 111
  • Publication Date: 2021
  • Doi Number: 10.1016/j.landusepol.2021.105739
  • Journal Name: LAND USE POLICY
  • Journal Indexes: Social Sciences Citation Index (SSCI), Scopus, PASCAL, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Environment Index, PAIS International, Political Science Complete, Pollution Abstracts, Public Affairs Index, Veterinary Science Database, DIALNET
  • Keywords: Land valuation, Land consolidation, K-means, K-medoids, Fuzzy C-means, Fuzzy Weighted differential evolution algorithm, MECHANIZATION, PRODUCTIVITY, ALGORITHM, SCHEMES, SUPPORT, NEED
  • Erciyes University Affiliated: Yes

Abstract

Land valuation is a comprehensive assessment process that aims to assign the agricultural value of all parcels in the land consolidation area, based on soil quality and land productivity (using a relative non-dimensional score). Thus, the land value represents a critical parameter that directly influences the monetary interests of landowners. This process should be managed in a reliable, correct, and fair manner. Furthermore, the traditional land valuation process is time-consuming and costly, and its results may be inconsistent because those who determine the value cannot take into account and compare the land valuation parameters required for all parcels. The solution to these deficiencies requires a new valuation approach. After land consolidation in the project area, the value of the existing parcels must be determined according to certain criteria in order to give to the enterprise land with the equal value to its previous land. In this study, a new land valuation model was developed with the help of clustering algorithms (K-means, K-medoids, Fuzzy C-means) and Weighted Differential Evolution, a heuristic optimization algorithm, using the most basic nine different parameters affecting the land value. The clustering method used in this model performs the valuation by clustering the parcels with common characteristics according to the parameter values. The method in which the cumulative sum of the distances of parcels to the cluster centers is the shortest exhibits the best clustering performance. In this study, the best clustering performance was obtained with the WDE-based clustering algorithm. When compared with the other algorithm results by mapping the classical valuation results, it was determined that the clustering method results evaluated the parcels more precisely. The study contributes to the literature in terms of including in the developed method parameters other than those used in the existing methods and determining the land value more precisely, fairly, and reliably with the help of heuristic algorithms.