Using Machine Learning to Detect Cyberbullying

Abstract

Abstract-Cyberbullying is the use of technology as a medium to bully someone. Although it has been an issue for many years, the recognition of its impact on young people has recently increased. Social networking sites provide a fertile medium for bullies, and teens and young adults who use these sites are vulnerable to attacks. Through machine learning, we can detect language patterns used by bullies and their victims, and develop rules to automatically detect cyberbullying content.The data we used for our project was collected from the website Formspring.me, a question-and-answer formatted website that contains a high percentage of bullying content. The data was labeled using a web service, Amazon's Mechanical Turk. We used the labeled data, in conjunction with machine learning techniques provided by the Weka tool kit, to train a computer to recognize bullying content. Both a C4.5 decision tree learner and an instance-based learner were able to identify the true positives with 78.5% accuracy.

Challenges in Modifying Existing Scales for Detecting Harassment in Individual Tweets

“…This approach would at least measure whether a user's content matches their personality. One study shows that at least one "real-life" personality trait often thought to be associated with aggression, narcissism (see [3,9,30,35]), persists in online environments; this is reflected in how people scoring high in narcissism conduct themselves on Facebook [8]. Even among the authors of this paper, however, there is disagreement about the utility of this approach given that people may behave differently in different online communities where norms of behavior are different [6,7].…”

“…Our plan was to use human coders, able to capture these subtleties, in order to generate training data for automated classifiers that could aid moderators and users in the detection of violent content. Our scale was designed to detect harassment [see , 35 for a report on binary [present/absent] harassment detection] and to indicate the specific type of harassment occurring.…”

“…In this model, it is either incumbent upon the victim or a thirdparty observer of harassment to report and/or manage the harassing content. Automated detection efforts have so far had limited success [see, e.g., , 15,35,41,45], but machine learning approaches hold promise [35,39,40].…”

Application Based on Fuzzy Logic to Detect and Prevent Cyberbullying Through Smartphones

“…Among these studies, there are some focused on the detection of cyberbullying through patterns in the analysed texts. For example, the paper [16] proposes the use of machine learning to detect cyberbullying. According to it, through automatic learning, the proposed tool can detect language patterns used by bullies and victims, and develop rules to automatically detect bullying content.…”

Technology Solutions to Combat Online Harassment

“…Previous work has applied anti-spam techniques like machine learning based text classification (Reynolds et al, 2011) to detecting harassing messages. However, existing public datasets are limited in size, with labels of varying quality.…”