초록 열기/닫기 버튼
-
Republic of Korea Army will reduce 11.4 thousand troops from 63.6 thousand to 52.2 thousand by 2022 according to the 'National Defense Reformation ＇12- ＇30 issued in 29, August, 2012. This states an increasing importance of the strength of 'Reserve forces and of 'High-tech forces. For this reason, the Ministry of National Defense is making effort to urge each subordinate forces to increase their mobilization response rate. To increase the mobilization response rate, it is important to predict the mobilization rate of soldiers able for mobilization in each subordinate unit and to manage differentiated services. Thus, I used data mining techniques to develop a classification model which classifies the soldiers able for mobilization into one of two mobilization groups using only restricted data of 'The National Defense Mobilization Information System'. The target variable for this model is a binary variable, whose value can be either 'a group of mobilization' or 'a group of non-mobilization'. I developed five classification models using Decision Tree, Neural Network, Support Vector Machine, Naive Bayesian and Multi Boost in data-mining techniques. Experimental results demonstrated that the highest performance model was the model of Decision Tree. We can expect that these classification models can be used in field units to effectively manage mobilization forces more efficiently.
키워드열기/닫기 버튼
Mobilization, Mobilization rate, Data-mining, Decison Tree, Neural Network
