Validation of Loss and Continuation Rate Models to Support Navy Community Management

Abstract

Continuation rate models are used to forecast the number of individuals available in a given time period by paygrade, years of service, and skill group. Forecasted manpower inventory, the number of individuals available in a given time period, are derived from continuation rates, which in turn inform the advancement and gains modules used within the Department of Defense. As such, the accuracy of continuation rates is critical to these related functions. The calculation of continuation rates can be based on business rules, statistical analysis, or a combination of both. Using enlisted personnel data from 2003-2011, we separately estimate and compare four different methodological approaches for a subset of enlisted skill groups. We validate each of the methods based on absolute error differences between forecast results and data observations. Of the four methodologies, the findings indicate significant differences in performance across models. Variation in model performance may be affected by sample size and factors not directly captured by the models, but having indirect effects on continuation rates, such as policies governing skill group contracts, promotions, bonuses, and or economic events. Based on forecasted results, we provide a detailed comparison of model performance.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 2014
Accession Number
ADA597724

Entities

People

  • Ping Ying Bellamy
  • Tanja F. Blackstone

Organizations

  • Navy Personnel Research, Studies, and Technology

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Accuracy
  • Bayesian Networks
  • Data Science
  • Databases
  • Department Of Defense
  • Enlisted Personnel
  • Information Science
  • Instructors
  • Management Personnel
  • Manpower
  • Military Research
  • Naval Personnel
  • Personnel Management
  • Statistical Analysis
  • Statistics
  • Students
  • War Colleges

Readers

  • Computational Fluid Dynamics (CFD)
  • Naval Personnel Management
  • Regression Analysis.