Certain Nuances in the Forecasting Models of the Social Phenomena

The paper considers the problem of constructing an adequate forecasting model for the large-scale and ambiguous social phenomena using crime dynamics in Russia as an example for analyzing. It focuses on the classical data analysis methods that have shown their high accuracy in forecasting the social phenomena. The paper argues that traditional statistical approaches complemented by a correlation analysis ensure the forecasting reliability without a need for the complex mathematical models and machine learning. This approach could be introduced to analyze various social phenomena and makes it possible to deeply study current trends and prospects in their development.