Estimation of housing sales prices by multiple linear regression analysis (MLRA) for the determination of economically effective earthquake based urban transformation locations

The aim of the study was to examine both physical and locational variables that might affect the house prices and estimate a price function using a multiple linear regression analysis and visualize the results by thematic maps. This study utilizes a large data set representing 390 home sales in Umraniye County in Istanbul/Turkey, which are researched with their physical and locational variables. The identification of the variable affecting the price of houses in the district with a population of approximately 600000 is expected to provide a significant contribution to the house marketing. Physical variables are chosen as the age, number of rooms, size, floor number of the apartment and the floor that the house is positioned in. Locational variables are chosen as to the nearest hospital, school, park and the police station distances to the house. The effects of the variables on the house prices are examined both by multiple linear regression analysis and hedonic regression analysis models, which are both stochastically analysis methods. The regression parameters found from both models leads almost to the same results.