Mobil Cbs Tabanlı Taşınmaz Değerleme Karar Destek Uygulaması Geliştirilmesi

Abstract

Real estate valuation is important in terms of taxation for government institutions, crediting and insurance sectors and determination of trading values for investors. Real estate valuation is carried out by licensed appraisers in Turkey and many other countries. The appraisal process starts with a request and ends with the expert's valuation and control by the auditor. With the advancements in information technologies these processes are being automated. In this study, it is aimed to reduce the time required by an expert for performing the precedent determination process in the market research by providing a mobile GIS platform in order to facilitate the value control made by the auditor before the final report. The application stores the previous evaluations for the estimation of real estate value in a geodatabase in conjunction with location and other important features. For each new real estate to be evaluated, some spatial attributes (distance to city center, hospitals, etc.) are automatically calculated by the developed software. An extensible architecture has been adopted to facilitate the integration of machine learning methods that will enable the automatic prediction of real estate valuation and is initially tested by integrating Inverse Distance Weighted (IDW), Ordinary Least Square (OLS) and Spatially Lagged Regration (SLR) methods. A self-learning regional valuation approach has been developed with the aim that each finalized value can be used for the next estimate. The developed approach and software is tested with the data of Mamak district in Ankara province and the results are discussed.