The article proposes an approach to the development of models for assessing the resilience of macroeconomic systems to the effects of exogenous shocks with an emphasis on the study of assessing the resilience of diagnostic classes. The relevance of the chosen research topic is explained by the fact that the development of the world economy takes place in the context of increasing globalization processes. In this economic environment, both positive and negative effects of these processes are formed. Due to the inability to control the impact of external destabilizing factors, there is a problem of assessing the resilience of the economy to "shocks" elements. Consideration of thematic, literature sources allowed to confirm the importance of the stated research and insufficient elaboration of the issues of assessing the stability of cluster formations. The aim of the article is to develop classification models that, based on hierarchical agglomerative methods of cluster analysis, iterative methods of cluster analysis, Kohonen neural networks, allow to analyze the stability of macroeconomic systems cluster formations, analyze the migration of elements from cluster to cluster systems to the action of exogenous shocks. The main objectives of the study were to develop models for classifying countries according to the level of resistance to exogenous "shocks" based on agglomeration methods of cluster analysis, iterative methods of cluster analysis, Kohonen neural networks; assessment of the classification quality, justification of the choice of the final breakdown; analysis of migration from cluster to cluster, assessment of structural dynamics. The obtained results allowed us to conclude that Kohonen neural networks provide an opportunity to obtain a better and more economically interpreted classification taking into account the models of crisis development in the element countries prone to migration from cluster to cluster. Analysis of the structural dynamics of clusters in the pre-crisis, crisis, post-crisis period shows a decrease in global stability, as there is a high proportion of countries with medium and low resistance to exogenous "shocks", as well as migration of many elements to the lower cluster in the post-crisis period. The analysis of cluster characteristics showed that for countries with a low level of resistance to exogenous "shocks" the critical subsystem is the financial security subsystem, which requires the transformation of protection mechanisms for financial "contagion". The obtained results can be used in systems of proactive crisis management.
