Modelling Fixated Discourse in Chats with Cyberpedophiles

A considerable variation in the length of sex-related lexical chains according to the nature of the corpus supports the belief that this could be a valuable feature in an automated pedophile detection system.

Abstract

The ability to detect deceptive statements in predatory communications can help in the identification of sexual predators, a type of deception that is recently attracting the attention of the research community. Due to the intention of a pedophile of hiding his/her true identity (name, age, gender and location) its detection is a challenge. According to previous research, fixated discourse is one of the main characteristics inherent to the language of online sexual predation. In this paper we approach this problem by computing sexrelated lexical chains spanning over the conversation. Our study shows a considerable variation in the length of sex-related lexical chains according to the nature of the corpus, which supports our belief that this could be a valuable feature in an automated pedophile detection system. 1