ВИКОРИСТАННЯ КАРТ КОХОНЕНА ДЛЯ АНАЛІЗУ РІВНЯ ІНФОРМАЦІЙНОЇ БЕЗПЕКИ КРАЇН З УРАХУВАННЯМ ЇХ РОЗВИТКУ

Abstract

The development of computer technology leads to the emergence of information threats in the country associated with the appearance of information wars and cyber terrorism. There is a need to analyze the level of information security of the country because of this reason. It is also necessary to take into account the level of country development. Two groups of indicators were selected for analysis – information security indexes and development indicators. Global Cybersecurity Index, National Cyber Security Index, ICT Development Index, Networked Readiness Index, Digital Development Level formed a group of information security indicators. Thirty-seven indicators of world development were included in another group. The dataset was generated for 159 world countries in 2018. A correlation analysis was carried out in this work to identify indicators with a close statistical relationship. As a result, 12 development indicators were selected. Non-linear normalization was also performed to bring the data into comparable values. Further research was carried out using an analytical platform Deductor Academic. Data were checked for quality, outliers, duplicates and inconsistencies. As a result, outliers were identified for three observations of the indicator “Life expectancy”, after which the data were replaced with probable values. Kohonen maps were constructed, taking into account different combinations of parameters, as a result of which the option with the lowest quantization errors and optimal hit density was chosen. Based on the results of the experiments, it was selected the method for determining the initial weights of neurons “From eigenvectors”, the neighborhood function “Stepped”, the error level for data recognition is less than 0.05, the size of the map is 24:18. As a result, seven clusters were obtained, which characterize groups of countries by the level of information security, taking into account the indicators of their development. Clusters “0” and “1” include countries with the highest level of development and information security. Group “2” characterizes countries with above-average indicators. The third cluster identifies countries with an average level of development and protection. Group “4” includes countries with indicators for which the degree is below average. The fifth cluster assesses countries with a low level of development and information security, and the sixth cluster characterizes the level of countries as very low.