?
Моделирование рынков жилой недвижимости крупнейших городов России
Increasing the efficiency of the real estate market is a large-scale national economic task, the effective solution of which depends on the possibilities of a reliable mass assessment and forecasting of the market value of residential properties. However, the existing methods and the mathematical models and computer programs created on their basis have a number of disadvantages. Firstly, mathematical models require frequent updating, since they do not take into account the influence of external macro- and mesoeconomic factors. Secondly, developed for one local market, they are not suitable for use in other regions and do not allow you to perform a comparative analysis and predict the development of markets. The objectives of the study are to create an intellectual system for assessing the market value of real estate that is self-adaptive to macro parameters and is scalable for use in various regions. The main research method is neural network modeling and data mining. The hypothesis of the study is that these goals can be achieved if neural networks are trained on examples of the behavior of markets in several cities at the same time, taking into account not only the construction and operational parameters of apartments, but also the geographical location of the regions, the degree of prestige and convenience of the location of the house, as well as the economic the state of each region, country and the whole world at the time of appraisal of the apartment. The advantage and scientific novelty of the created intellectual system is the property of its self-adaptation to space and time. The functioning of the intellectual system is demonstrated on the example of scenario forecasts for the development of real estate markets in Moscow, St. Petersburg and Yekaterinburg, which made it possible to identify a number of significant patterns in local real estate markets located in different regions. Conclusions from the results of the performed computer experiments, the developed model and the computer program that implements it are in demand by government agencies involved in the regulation of urban real estate markets, property taxation, as well as strategic planning of construction organizations.