EXTRACTING MAJOR TOPICS FROM SURVEY TEXT RESPONSES USING NATURAL LANGUAGE PROCESSING
Loading...
Authors
Layug, Christine
Subjects
text analysis
SURVEY
Naval Inspector General
Defense Equal Opportunity Management Institute
DEOMI
Organizational Climate Survey
DEOCS
SURVEY
Naval Inspector General
Defense Equal Opportunity Management Institute
DEOMI
Organizational Climate Survey
DEOCS
Advisors
Whitaker, Lyn R.
Date of Issue
2018-09
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
In this thesis, we enhance the comment analysis approach of Cairoli’s 2017 Naval Postgraduate School thesis, to help the fleet analyze comment responses in Department of Defense surveys and utilize the results to make important decisions. This methodology automates applying descriptive labels to a comment and then uses those labels to categorize comments into a small set of meaningful prevalent topics. We apply this methodology to comments from two recent surveys: a command climate survey as well as an investigation survey looking into the recent increase of physiological episodes experienced by T-45 and F/A-18 aircrews. When applying novel approaches to different data, unexpected matters emerge. These matters shed light on areas of the approach that may need expansion or modification. Motivated by our analysis of text comments from two very different Navy surveys, we extend Cairoli’s approach in four ways. Our modifications lead to a generalized model; an approach independent of the need to acquire and preprocess an external reference corpus; more automation of the topic discovery process; and an added element that allows a comment to have more than one topic.
Type
Thesis
Description
Series/Report No
Department
Operations Research (OR)
Organization
Identifiers
NPS Report Number
Sponsors
Funder
Format
Citation
Distribution Statement
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.