Descriptive Models of Monthly Officer Attrition

摘要

Logistic Regression modelling techniques are applied to the problem of describingMarine Corps officer monthly attrition based upon such factors as grade, military occupation specialty, promotion zone, season and gender. The monthly unemployment rate is considered as a useful covariate. The monthly losses are rather rare events and the ability to support the suggested models is not firm. Some exploratory computations are performed in order to examine the effect of using these models in conjunction with previously established annual rate generating models. Manpower models, Rate generation, Logistics regression.