Accession Number : ADA114770
Title : Aid Loss Probability System for Officers (AID-O). Volume 1. Executive Summary.
Corporate Author : GENERAL ELECTRIC CO WASHINGTON DC TEMPO CENTER FOR ADVANCED STUDIES
Personal Author(s) : Powell,Ken R ; Kim,Kwan H ; Schilling,Terry L
PDF Url : ADA114770
Report Date : 15 Sep 1974
Pagination or Media Count : 61
Abstract : The AID-O system is capable of generating consistent and reliable loss probabilities for current and future generations of officer projection models. The core of the system is a statistical technique developed at the University of Michigan, known as the Automatic Interaction Detector (AID) routine. This technique groups the actual officer population into subpopulations identified by personnel attributes which are most predictive of individual losses in each subpopulation. A loss probability is developed for each individual who belongs to a specific subpopulation. Any desired loss matrix format then can be generated by summing the probabilities over the appropriate individual attributes and dividing by the number of occurrences in each cell. This flexibility provides the analyst with the capability for creating a variety of loss matrices with consistent loss probabilities. The AID-O system allows the analyst to identify the loss probabilities associated with various officer subpopulations which will be affected by alternative policy decisions. The historical rates for these subpopulations then can be subjectively adjusted and a new set of loss matrices derived for each policy alternative. Finally, using Central Integrating Model-Officers (CIM-O) or manual modeling methods, the projected impact of the policy alternatives on the officer population can be evaluated. The AID-O system provides the analyst with a clear identification of the attributes associated with the various levels of loss behavior, as well as reliable means to generate accurate and consistent loss probabilities.
Descriptors : *Models, *Manpower, *Army planning, *Army personnel, *Officer personnel, Army, Policies, Predictions, Losses, Probability, Variables, Retirement(Personnel)
Subject Categories : Administration and Management
Personnel Management and Labor Relations
Logistics, Military Facilities and Supplies
Distribution Statement : APPROVED FOR PUBLIC RELEASE