DEVELOPMENT AND RESEARCH OF MULTISIZED REGRESSIVE MODELS OF SECOND HAND MARKET OF RESIDENTIAL REAL ESTATE

Purpose of the performed research were: to define the content of the meaningful price forming factors, which have significant influence on the cost of real estate at the second hand market; test of hypothesis about the connection of the cost for residential real estate and dates of its exposition at the market and definition of the fact how close this connection is. Methods. Development and research of multisized regressive models of the second hand residential real estate are performed at the example of the offers about flat selling in the panel block of flats of the 135th series, located in the microdistrict Universitetskii of the city of Irkutsk. Taking into account not linear character of the dependence of cost indexes on the majority of price forming factors that exist at the market of real estate, we used linear in parameters (additive) multisized regressive models. Results and their discussions. As a result of the performed research we built a multisized regressive model. The combination of statistic indexes of the model allows to consider it as the accounting main part of the price forming factors that significantly influence at the considered segment of real estate market. Conclusions. The content of the meaningful price forming factors that significantly influence the cost of real estate at the considered segment of the second hand market of residential real estate: general square of the room, floor, inside finishing and its condition, presence of a balcony and its glazing, room planning, presence of furniture, age of the building, location of the building in relation to transport ways distance to the public station, to the hospital, presence of the food shop in the building, date of exposition of the object at the market. Connection of the prices of the offer and dates of the exposition at the market is high; the pointed out factor should be taken into account during the process of building of multisized regressive models in order to increase the level of their approximation