Регрессионно матричная модель многомерных экономических процессов

Abstract

Purpose: building of complex models, providing the possibility of multivariate forecast calculations of socio-economic development of the region. Discussion: the multidimensionality and multiplicity are the main problems that must be overcome in the forecast,calculations,of the expected socio-economic status of the region. The model, which could solve both problems, is not currently known. The development of such model, according to the authors is possible on the basis of combining the econometric approach with the ideas of multidimensional prediction based on the matrix predictor. The important point in the implementation of this combined approach is a method by which diverse models can be successfully combined, giving the result of combining a focus on getting the desired result. This method was obtained by a special representation of the autoregressive diagram of modelling. The multidimensionality and multiplicity in the final version are ensured by a set of three models, two of which, as a subsidiary, provide the completeness of information support of the combined model. Results: the method of obtaining regression-matrix model allowing multiple calculations of a multidimensional processes in the analytical form is described in detail.