географически взвешенная регрессия

Introduction. Detection and measurement of interdependencies in the housing market is one of the key issues examined econometric methods. Compared with traditional methods of geographically weighted regression extends the understanding of how the units belonging to the set of specific geographical coordinates affect the relationship between the covariates and the price of real estate. In this regard, the aim of this study was to analyze the spatial differences in the price of one-bedroom apartments presented in the secondary housing market of Orenburg. Methods. We used the method of cluster analysis, graphical method, analysis of variance, the classical regression model and geographically weighted regression. Results. Parameter estimation of the global (general) model by least squares and geographically weighted regression, has shown that SMT has a better fit, and is proof of the spatial differentiation of the regression coefficients. Conclusions. When modeling the price one-room apartment to be preferred geographically weighted regression, since it is estimated regression coefficients for each object combination and therefore recognized geographic differences in the dependencies, it is difficult to display the total regression equation.