Опыт оценки рисков крупномасштабной вооруженной политической дестабилизации в странах Африки с использованием методов машинного обучения

История и современность. 2024. № 3 (53). С. 18–44.

Медведев И. А., Ustyuzhanin V., Zinkina J. V., Chernomorchenko I., Korotayev A.

The article presents an assessment of the risks of large-scale political destabilization/civil wars in African countries using machine learning techniques. The main focus is on the application of algorithms, in particular the CatBoost model, to the analysis of a wide range of data from interdisciplinary sources (economic, social and political). A distinction is made between stable countries and countries that have experienced or are experiencing instability. The highest risks of political destabilization are predicted for large, poor countries with natural resources, high population growth rates, and low levels of urbanization and education. The most stable countries are confirmed to be small and island states with homogeneous populations, as well as the countries of southern Africa, which are characterized by low birth rates, high urbanization and education levels compared to other regions of the African continent. Particular attention is paid to the importance of demographic factors, confirming the relevance of the “youth bulge” theory in the analysis of large-scale political destabilization in Africa. It is argued that the presence of easily accessible natural resources, such as gold or oil, significantly increases the risks of large-scale political destabilization in countries with weak political institutions.

Keywords: природные ресурсы, политические институты, Африка, база данных, database, Africa, риски дестабилизации, political institutions, дестабилизация, destabilization, natural resources, youth bulges, destabilization risks, молодежные бугры