A Production Early Warning System (PEWS) Model Which Predicts Future USAREC Mission Accomplishment

Abstract

This thesis develops a framework for a statistical Production Early Warning System (PEWS) model which predicts the United States Army Recruiting Command's contract production. Model predictions are based on the initial Armed Forces Qualification Test (AFQT) taken by applicants over the past two years, the number of applicants expected to take the AFQT throughout the projection period, the historical probability that an applicant will sign a contract, and the distribution of time from when applicants take the AFQT until they sign a contract. Model parameters are based on the last five years of historical testing and contracting data. Yearly, seasonal, and monthly trends are incorporated by analyzing historical data using semi-monthly segments split on the 15th of the month. The model predicts contract production overall and for seven separate mission box categories. Performance of the model is measured by subtracting the number of actual contracts from the number of predicted contracts, and dividing by the number of actual contracts for FY 1993 time periods. The model's accuracy is greatly reduced because the testing data base does not include applicants who took the AFQT as part of a batch test group.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1994
Accession Number
ADA289750

Entities

People

  • Scott G. Roesler

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Accuracy
  • Basic Training
  • Computer Programs
  • Computers
  • Contracts
  • Databases
  • Early Warning Systems
  • Lists (Data Structures)
  • Organizational Structure
  • Personnel Management
  • Probability
  • Recruiting
  • Spreadsheet Software
  • Students
  • Training
  • United States
  • Warning Systems

Readers

  • Computational Modeling and Simulation
  • Naval Personnel Management