Accession Number:



Time Series Models for Predicting Monthly Losses of Air Force Enlisted Personnel

Descriptive Note:

Research rept.

Corporate Author:


Report Date:


Pagination or Media Count:



The Short-Term Aggregate Inventory Projection Model SAM is one component of the Enlisted Force Management System EFMS. SAM makes monthly projections for the rest of the current fiscal year of the aggregate force the total enlisted force across all specialties. It can be used to analyze the total size, grade composition, and budget cost of the enlisted force during a fiscal year. It supports planning of management actions to achieve user specified end-of-year force levels known as end strengths and user specified end-of-year grade levels known as grade strengths. SAM contains five modules -- Module P Preprocessor, Module 1 Separation Projection, Module 2 Inventory and Cost Projection, Module 3 Computer-Aided Design, and Module 4 Plan Comparison. Module 1 predicts policy-free monthly losses and reenlistments of Air Force enlisted personnel for the rest of the current fiscal year. Policy-free means that the predictions assume zero early releases and zero early reenlistments caused by actions of enlisted force managers. Time series models are one way of predicting the separations required from Module 1. In general, five distinct types of separations must be predicted Losses on or close to the expiration of an airmans term of service ETS losses, Losses during the term attrition losses, Losses when an airman is on extension status, Retirements, and Reenlistments. Different models were used to predict each of these types of losses for each term of an airmans career. To find the model that fitted the historical data best and was likely to produce the best predictions, four time series modeling approaches were tried Constant rate, Regression, Autoregressive, and Straight line running average.

Subject Categories:

  • Personnel Management and Labor Relations
  • Computer Programming and Software
  • Military Forces and Organizations

Distribution Statement: