Akademik Veri Yönetim Sistemi
Journal of Housing and the Built Environment, cilt.38, sa.3, ss.2005-2027, 2023 (SSCI)
In residential real estate, location quality within the neighborhood has considered a very important characteristic. On account of a limited number of studies on novel evaluation methods in real estate literature, an Adaptive-Neuro-Fuzzy-Inference-System Grading Model (ANFISGM) which combines human knowledge-based FIS with optimization skills of ANN is proposed. By 4 location parameters, Distance to Open Space (DtOS), Parcel Size (PS), Excellence of View (EoV), and Distance to Parking (DtP), the model has been applied to grading the location quality of 27 detached properties within the campus region of Kocaeli University. Performance of the proposed model is compared with the Standard Grading Method (SGM), and Fuzzy Grading Model (FGM), which were developed considering the quality of property location (QoPL) within the neighborhood. In this study, the ANFIS model is proposed as the first contribution to the real estate valuations made in terms of location quality in the neighborhood. Since it was not very clear in previous studies, the individual and resultant effects of the parameters on the score are examined and interpreted by correlation surfaces. The lack of nonlinear interpolation, which reduces the sensitivity of traditional and fuzzy methods and causes them to assign the same value throughout the transition regions, is eliminated by the proposed ANFIS method. Hereby, ANFISGM declares an accurate and practical grading model to estimate the value of the residential real estate considering property location within the neighborhood. This paper expects that ANFISGM will contribute to the appraisal process within a minimum of detail and a limited period, especially for tax purposes.