Using machine learning for automated role identification in cyberbullying

Date of Publication

2014

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 Cheng

Abstract/Summary

Bullying has been an old problem that experts believe does not cease to exist as people grow older (Krantz, 2012). With the advent of computer technology, bullying has also evolved from being a physical experience to a virtual experience, now widely known as cyberbullying. Since traditional bullying involves the participation of different roles, the proponent speculates that the same roles are also present in cyberbullying. Existing researches included determining texts which contained cyberbullying, while a few involved the use of roles in determining whether bullying occured or not. The problem is that these models are created for classifying texts written in English, and thus cannot be used in the local context. By using social media sites as sources for model training, the research created a support vector machine (SVM) based model for detecting bullying through the use of roles by using word features which were selected using the TF-IDF algorithm combined with the use of weights. 10-fold cross validation showed an accuracy of 59.7% for using 171 unique word features with a Kappa statistic of only 42.3%.

Abstract Format

html

Language

English

Format

Electronic

Accession Number

CDTG005722

Shelf Location

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

Physical Description

1 computer disc ; 4 3/4 in.

This document is currently not available here.

Share

COinS