Analysis of the effectiveness of the discriminant function in solving the problem of country classification basedon a set of indicators

Introduction. The relevance of studying discriminant analysis in solving economic problems is due to its ability to improve decision-making, forecasting, risk management, and adaptation to changes. In a dynamic and competitive economic environment, this method becomes an important tool for analyzing and optimizing various business processes. The study is aimed at investigating the comparative effectiveness of the method for constructing a discriminant function in solving the problem of classifying countries based on a set of indicators. Materials and methods. Publications in peer-reviewed journals on economics, statistics, and data analytics were used, as well as materials from conferences dedicated to new approaches in discriminant analysis. Methods included theoretical literature analysis, modeling for forecasting and classification. The research task involved analyzing seven macroeconomic indicators of 25 countries (data from open sources). The goal was to identify key features, study internal discrimination, and construct a linear hyperplane for classifying countries based on these features. Results. The results of country classification using discriminant analysis are presented. Factors determining country discrimination were identified. The weight and significance of each studied macroeconomic indicator in discrimination were analyzed. The discriminant function and discriminant constant were determined, allowing the group of countries to be divided into two classes, each with its own within-group correlation distribution. New objects were classified using the constructed discriminant function. Conclusion. Discriminant functions were constructed to classify countries based on seven macroeconomic indicators, taking into account their weight and significance. Differences in within-group correlation allow countries to be divided into two classes. Key factors identified include GDP, exports, imports, foreign direct investment, and population size. The results demonstrate high model robustness. In the context of technological development and Big Data, discriminant analysis can be enhanced through machine learning and AI, expanding its application in finance, marketing, healthcare, and other fields.