Open AccessThe influence analysis of criteria of comparables’ selection on the accuracy of the property value estimation
Author(s) -
Jacek Zyga
Publication year - 2020
Publication title -
reports on geodesy and geoinformatics
Language(s) - English
Resource type - Journals
eISSN - 2391-8365
pISSN - 2391-8152
DOI - 10.2478/rgg-2020-0002
Subject(s) - mathematics , similarity (geometry) , euclidean distance , property (philosophy) , selection (genetic algorithm) , value (mathematics) , set (abstract data type) , statistics , basis (linear algebra) , relation (database) , variable (mathematics) , econometrics , mathematical optimization , computer science , data mining , artificial intelligence , mathematical analysis , philosophy , geometry , epistemology , image (mathematics) , programming language
The subject of the considerations presented in the article is the question of the criteria according to which the comparables, in the market value estimating process, should be selected. As the most important in the selection of comparative properties, the factor of similarity in relation to the subject property (measured by the Euclidean distance) was considered. As the key issue, the similarity assessment criterion and the influence of the adopted critical value of this criterion on the accuracy of estimates were chosen. The analysis of the above was carried out with taking into account the role of the significance of independent variables (measured by their correlation with the vector of the dependent variable). The results of the simulation tests carried out in the variants set by the criteria adopted were presented. On this basis, it has been shown that there is a potential most relevant solution in the collection of obtainable estimation results. This solution corresponds with the smallest sum of the differences of model prices (accepted as known) and corresponding estimates. The found minimum occurs for a specific layout of the above criterion values only.