Dissertation/Thesis Abstract

Matching Students with Support Services through a Constrained Linear Optimization Model
by Gonzalez, Diana, M.S., California State University, Long Beach, 2018, 64; 10828179
Abstract (Summary)

College students have several on-campus resources available to increase academic success and generally do not have an effective way for selecting the right one. We were inspired to create a decision aid to help students pick their resources due to the success of decision aids used in the medical field. An instrument to match students with the most appropriate support service by studying existing data was constructed by analyzing roughly 350 student responses to questions from the Ruffalo Noel Levitz survey and a questionnaire, totaling 108 questions, both administered by TRiO Student Support Programs at CSULB. One of the goals of this project is to make a shorter survey that would capture the same information as the original.

To that end, Principal Component Analysis was performed on the existing data to determine the number of categories needed for the survey. Factor Analysis was then used to select the representative question from each category. The new survey remains valid since the Ruffalo Noel Levitz survey was validated in the first place.

To develop our model for matching students with the most appropriate support services, we asked students in the Early Start Mathematics Program at California State University, Long Beach to answer the new survey and rank their preferred resources; we collected 300 data points, 162 of which are usable. The ranked resources were used as ground truth to create a training set. A constrained least-squares optimization model was proposed to match the survey responses with the ranked resources. The solution to the least-squares problem, solved via the method of Langrangian-Multipliers, gave us a way to match survey responses to resources. k-Fold Cross Validation was used to validate the accuracy of the method by using the ground truth gathered from the ranked resources.

Indexing (document details)
Advisor: Chang, Jen-Mei
Commitee: Lee, Chung-Min, Zhou, Tianni
School: California State University, Long Beach
Department: Mathematics and Statistics
School Location: United States -- California
Source: MAI 58/02M(E), Masters Abstracts International
Subjects: Applied Mathematics, Higher education
Keywords: Math modeling, Principal component analysis, Student resources, Student support services, Survey analysis
Publication Number: 10828179
ISBN: 9780438568150
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy