Estimating the market value of residential buildings with artificial neural networks method: Düzce sample

Housing is real estate that meets people's housing needs. Nowadays, valuation services are needed to generate financial resources in production, marketing and sales stages. Accurate assessment of the fair values of the houses will be possible with the correct evaluation of the parameters. Different values emerge with the emergence of personal opinions in the fast and accurate conclusion of the valuation process. Inconsistent prices between peer real estates create difficulties in valuing to experts. In this study, artificial neural networks, one of the artificial intelligence methods, are used to evaluate the fair values of the houses quickly and accurately. Artificial neural networks (ANN) are a prediction method and perform very successfully when it is difficult to obtain output data depending on the input data. Within the scope of the study, a model based on ANN and Regression Analysis (RA) has been developed for the purpose of estimating the market values of 150 houses valued by the real estate appraiser. The data set, which was created by using the existing data, was entered as data into artificial neural networks structured in single and multi-layered, feed-forward, consultant learning features. As the input vector, 22 criteria of the houses were used. The fair values found by the precedent comparison method of the houses were used as the output vector. The performance of this method in terms of market value, duration and proximity to reality were investigated by utilizing the features of learning, information storage and generalization. The solutions calculated by ANN, the valuation value, the RA method were compared and the error rates of the market value estimates were evaluated. According to the results obtained by ANN method, value estimation with 3.58% error rate could be realized. According to the regression analysis data, (the error rate 59.50%) is more realistic and applicable quality. At the end of the study it was seen that ANN modeling method can be used successfully in the pre-estimation phase of the market values of the houses.