APPLYING MACHINE LEARNING AND SENTIMENT ANALYSIS TO ASSESS THE HEALTH OF MARINE CORPS CULTURE
Loading...
Authors
Irby, Deondra I.
Subjects
natural language processing
NLP
sentiment analysis
topic extraction
BERTopic
topic modeling
Marine Corps culture
Command Climate Survey
DEOCS
large language models
Llama 2
Orange3
NLP
sentiment analysis
topic extraction
BERTopic
topic modeling
Marine Corps culture
Command Climate Survey
DEOCS
large language models
Llama 2
Orange3
Advisors
MacKinnon, Douglas J.
Zhao, Ying
Date of Issue
2024-06
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
The Marine Corps' commitment to maintaining a robust Esprit de Corps faces challenges from persistent and pressing issues for commanders like military suicides, mental health concerns, sexual harassment, assault, and equal opportunity complaints. This thesis explores the capability and usefulness of Large Language Models (LLMs) to enhance and improve command climate analysis within the Marine Corps by harnessing Natural Language Processing (NLP) to examine free-text responses in the Defense Equal Opportunity Climate Survey (DEOCS). This research rigorously compares the performance of LLMs against traditional machine learning models with regard to sentiment analysis and topic extraction. Our findings reveal that LLMs can provide more accurate and nuanced insights into free text responses of command climate surveys. LLMs also have the potential to expedite the analysis process and provide trend analysis on surveys aggregated over time and throughout commands using these methods along with the proper computational power. These advantages are critical for identifying key areas of concern and fostering troop welfare and retention across the Department of Defense. The deployment of these advanced NLP techniques found in LLMs can represent a significant leap forward in identifying operational efficiencies and can improve strategic decision-making within military contexts.
Type
Thesis
Description
Series/Report No
Department
Organization
Identifiers
NPS Report Number
Sponsors
Funder
Format
Citation
Distribution Statement
Distribution Statement A. Approved for public release: Distribution is unlimited.
Rights
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.