통계적 클러스터링을 활용한 새로운 신병 분류 방법 제안

정재범, Jae-Young Lee · 대한산업공학회 추계학술대회 논문집 · 2009

In this paper, we introduced a very effective Recruits Troop Classification(RTC) using a Statistical Clustering. Each recruit has a multi-variable reflecting individual characteristics. Current Army RTCs method is random classifying to input random number on web system. This random classification is secured fairness and clearness, but a lot of concerned recruits can be concentrated to a specific Troop. Therefore, wo propose new classifying method to make all troops have similar quality level for recruit members in each troop. To build the method, at first, we analyze structure of variable using PCA and choose k-means methods k as a clustering technique. And then execute random number classification in each cluster after making k clusters throng using clustering technique. To show the superiority of RTC method, we compare the RTC results to the one from current Army classification method, and the RTC results are clearly better than current Army method.

Cited by 0 works.

Authors