Dissertation/Thesis Abstract

Predicting Mental State Using Social Media Posts
by Seshadri, Subramaniam, M.S., California State University, Long Beach, 2019, 56; 13863615
Abstract (Summary)

With its increasing use, social media has become a ubiquitous aspect of modern society. Social media users belong to a broad spectrum of age, ranging from teenage kids to old adults. The primary aim of this thesis is to compare and analyze two different approaches, Glove and Word2vec, utilized for the modeling of language within social media posts to assess the mental state of users. These two approaches along with supervised learning indicate whether or not a social media user demonstrates textual usage risk-factors of depression. This thesis analyzes two popular machine learning approaches for classifying users’ linguistic tendencies based on these language modeling approaches and tests the effectiveness of diagnosing individuals for depression, based on social network textual content. This thesis highlights the differences between the two modeling approaches and explores the causality that results in similar conclusions between both methodologies. Utilizing two different algorithms, Random Forest and Support Vector Machines, social media users were classified for exhibiting linguistic risk-factors of depression. This thesis compares the results between both modeling and machine learning algorithms.

Indexing (document details)
Advisor: Ebert, Todd
Commitee: Tankelevich, Roman, Morales-Ponce, Oscar
School: California State University, Long Beach
Department: Computer Engineering and Computer Science
School Location: United States -- California
Source: MAI 81/3(E), Masters Abstracts International
Subjects: Artificial intelligence, Computer science, Sociolinguistics
Keywords: Deep learning, Machine learning, Random forest, Sentiment analysis, Support vector machines, Word vectorization
Publication Number: 13863615
ISBN: 9781088337325
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy