An Analysis of Modeling Success in Explosive Ordnance Disposal Training

Abstract

This thesis is a follow-on study to the Master of Business Administration (MBA) project, Grade Point Average as a Predictor of Success at Explosive Ordnance Disposal Training, completed in December 2009 by Lieutenant Sarah Turse and this author. The purpose of this thesis is to analyze, develop and provide a more accurate student graduation prediction model than the current model in place at Naval School Explosive Ordnance Disposal (NAVSCOLEOD). The school's current model was produced five year ago using ordinary linear regression. This outdated model was compared to the new model generated in this study using statistical techniques such as receiver operating characteristic (ROC) curves and the Hosmer-Lemeshow test. Our analysis finds that the student's branch of service, GPA, and the division in which the student failed each significantly impact predicting a student's future. We also find that the interaction between GPA and division also significantly impacts the prediction. Finally, we conclude that using a logistic regression instead a linear regression captures the binary output (graduated or did not graduate) better. Our improved model increases the prediction probability by roughly 2 percent using student data from 2004 to 2008.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2010
Accession Number
ADA518689

Entities

People

  • Trevor J. Ritland

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Biomedical
  • Counter IED
  • Human Systems

DTIC Thesaurus Topics

  • Air Force
  • Attrition
  • Business Administration
  • Education
  • Explosive Devices
  • Explosive Ordnance Disposal
  • Explosives
  • Improvised Explosive Devices
  • Instructors
  • Munitions
  • Students
  • Training
  • United States
  • United States Naval Academy
  • Warfare
  • Weapons
  • Weapons Of Mass Destruction

Readers

  • Aviation Safety Risk Assessment.
  • Computational Modeling and Simulation
  • STEM Education

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference