Detecting Predatory Behavior in Game Chats

Abstract

While games are a popular social media for children, there is a real risk that these children are exposed to potential sexual assault. A number of studies have already addressed this issue, however, the data used in previous research did not properly represent the real chats found in multiplayer online games. To address this issue, we obtained real chat data from MovieStarPlanet, a massively multiplayer online game for children. The research described in this paper aimed to detect predatory behaviours in the chats using machine learning methods. In order to achieve a high accuracy on this task, extensive preprocessing was necessary. We describe three different strategies for data selection and preprocessing, and extensively compare the performance of different learning algorithms on the different datasets and features.

A Technology Acceptance Model-Based Analytics for Online Mobile Games Using Machine Learning Techniques

“…The popularity of online mobile games is increasing as the adaptability of technology access by young people is high [27]. A notable number of techniques have been applied to evaluate the behavior patterns.…”

Early Detection of Sexual Predators in Chats

“…To our knowledge, the only work using real grooming chats is Cheong et al (2015) who used chats extracted from MovieStarPlanet, a massively multiplayer online game for children. Unfortunately, this corpus is not publicly available.…”

Identificação de predadores sexuais brasileiros em conversas textuais na internet por meio de aprendizagem de máquina

“…Na primeira iniciativa documentada para estudo do comportamento predatório na comunicação em jogos online [Cheong and Jensen 2015], o jogo MovieStartPlanet, cujo público-alvo são crianças e adolescentes entre 8 e 15 anos, disponibilizou três conjuntos de dados contendo a comunicação escrita de diferentes jogadores: mensagens de status, comentários em vídeos e postagens em fóruns, além de conversas públicas e privadas realizadas dentro da plataforma. Em dois dos conjuntos, foram considerados apenas conteúdos não predatórios e o terceiro conjunto de dados foi criado a partir de conversas realizadas por 59 predadores sexuais.…”