REGRESSION MODELS FOR ESTIMATING THE MARKET VALUE OF REAL ESTATE IN ST. PETERSBURG

The work is devoted to the urgent problem of obtaining an objective mass assessment of the market value of residential real estate. The purpose of the study is to use the example of the dynamically developing market of one-room flats in St. Petersburg to obtain an optimal, in terms of accuracy, regression model of real estate assessment. Regression analysis methods had been used to analyze the influence of various price-forming factors, including the characteristics of the location of real estate and the level of comfort of living. Four optimized models of real estate assessment have been obtained. The models differ in the factors characterizing the level of comfort of living, as well as in their structure: additive and multiplicative. Testing of the obtained models based on a control representative sample made it possible to select a model with maximum estimation accuracy (MedAPE=5,2%) and recommend it for practical use. As an example of the practical use of the model, the identification of undervalued and overvalued flats based on the obtained interval estimates of the average value of the real estate have been considered.