Abstract:

Since there are hundreds of parameters that affect the real value of the houses to be sold, determining the fair value is a very important issue. In this study, models are developed and the results are compared with both Multiple Regression Analysis (MRA), which is one of the statistical methods, and Artificial Neural Network (ANN), which is one of the artificial intelligence methods, in order to estimate the fair values of houses that depend on multiple parameters quickly and accurately. For this purpose, a total of 149 houses for sale in different neighborhoods of Kecioren district of Ankara, which were advertised on an e-commerce website where real estate sales are made in Turkey, were taken into consideration to create the MRA and ANN models. The 6 most effective parameters among hundreds of parameters in determining the fair value of a house were digitized and MRA and ANN models were established. Considering these data, it is seen that this selected ANN method gives more successful results compared to MRA.

Keywords: