Machine learning deployment for arms dynamics pattern recognition in Southeast Asia region

Abstract

Finding the most significant determinant variable of arms dynamic is highly required due to strategic policies formulations and power mapping for academics and policy makers. Machine learning is still new or under-discussed among the study of politics and international relations. Existing literature have much focus on using advanced quantitative methods by applying various types of regression analysis. This study analyzed the arms dynamic in Southeast Asia countries along with its some strategic partners such as United States, China, Russia, South Korea, and Japan by using 'Decision Tree' of machine learning algorithm. This study conducted a machine learning analysis on 55 variable items which is classified into 8 classes of variables videlicet defense budget, arms trade exports, arms trade imports, political posture, economic posture, security posture and defense priority, national capability, and direct contact,. The results suggest three findings: (1) state who perceives maritime as strategic drivers and forces will seek more power for its maritime defense posture which is translated to defense budget, (2) big size countries tend to be an arms exporter country, and (3) state's energy dependence often leads to a higher volume of arms transfers between countries.