Characteristics of the impact of various scenarios on the development of tourist infrastructure

This article will demonstrate a mechanism that tracks the dynamics of tourist flows between the regions of Russia under the influence of environmental factors and, as a result, tracks the dynamics of destination income/expenses from tourism activities and the level of infrastructure development. This algorithm is based on machine learning methods. In the course of the article, the importance of building two types of models will be proved. Models of the first type predict the total tourist flow, measured by the number of overnight stays and the amount of income of collective accommodation facilities (DAC), to a region with specified characteristics. Models of the second type predict cross-flows between regions, while the departure region and the arrival region are described by a different set of features. The article notes that the use of two types of models is due to the following reasons. Firstly, it is worth noting that the world experience of forecasting tourist flows is represented in most cases by models of forecasting tourist arrivals without analyzing the geographical structure of demand. Moreover, the presented models, with rare exceptions, do not assume the use of environmental factors and the level of development of tourist infrastructure as explanatory variables. Thus, the scientific task set out in this article is a new one. Secondly, there are significant differences in the methods of collecting statistical information about the tourist flow.