Title

Utilizing contextual information from Tweets as parameters for community detection input graphs

Author

Alron Jan Lam

Date of Publication

2017

Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science

College

College of Computer Studies

Department/Unit

Computer Science

Thesis Adviser

Charibeth K. Cheng

Defense Panel Chair

Arnulfo P. Azcarraga

Defense Panel Member

Charibeth K. Cheng
Nelson M. Marcos
Merlin Teodosia C. Suarez

Abstract/Summary

Twitter, as a microblogging platform, has become an avenue for people to voice out their opinions online. However, to effectively utilize this source of information, the massive amount of Tweets must first be processed to quickly obtain insights. One such way to achieve this is through community detection. Through this technique, Twitter users can be grouped into different types of communities such as those who interact a lot, or those who have similar sentiments about certain topics. However, most works do not utilize tweet content and simply use directly available information like Twitter follows. Hence, this work explores the utilization of hashtags and sentiment analysis (taking into account conversational context) as parameters in the input graph for community detection. Though the modularity score does not indicate much effect, an evaluation of topic model similarity of the communities tweets through word overlap and normalized pointwise mutual information show that differing contextual information and graph construction schemes can produce different insights. It is not necessary that one is better than the other, but rather, these are multiple approaches to getting insights for the end-users goals.

Abstract Format

html

Language

English

Format

Electronic

Accession Number

CDTG007291

Shelf Location

Archives, The Learning Commons, 12F Henry Sy Sr. Hall

Physical Description

1 computer disc ; 4 3/4 in.

Keywords

Online social networks; Microblogs; Social networks

This document is currently not available here.

Share

COinS