A House Valuation with Multiple Regression Analysis and Artificial Neural Networks

Abstract

Valuation, in its simplest form, is the determination of the amount that a property will be processed at a certain date. Valuation can be done for many purposes. These; can be listed as buying and selling, transfer, tax assessment, expropriation, inheritance distribution, investment, financing and credit. There are various methods of valuation. These methods are examined under 3 main groups as traditional, statistical and modern valuation methods. The aim of the article is to provide an overview of regression analysis, one of the statistical valuation methods, and artificial neural networks, one of the modern valuation methods, and to compare the accuracy values. Matlab software was used for artificial neural network modeling and Minitab software was used for regression analysis. The accuracies of the obtained values were determined by the average absolute percent error (MAPE) formula.