Una Aproximación para la detección de contenido sexual en conversaciones digitales

Abstract

INTRODUCTION. The internet has become more accessible using ICTs (information and communication technologies), the use of social networks has grown and has generated the exposure of minors to crimes such as cyberbullying, sexting, pederasty, among others. OBJECTIVE. Perform an analysis of conversations with pederast content and establish a prototypical structure used by criminals METHOD. Using the conversations obtained from Perverted-Justice, a grammar checker known as LanguageTool and a package of linguistic and text extraction tools such as Linguakit. These conversations could be tagged, categorized, and processed in different stages to obtain a corpus. RESULTS. Verbs, nouns, adjectives and adverbs most used by criminals were found. In addition, a scheme of interventions that provides a preliminary profile of the victimizer. DISCUSSION AND CONCLUSIONS. The analysis of conversations in the English language by pedophiles and pederasts on the Internet shows the lexicon that reflects the names and nouns that are used to obtain information such as photographs and type of clothing. In addition, adjectives are used to flatter the victim and thus gain their trust, obtaining the information easier. The importance of creating a corpus in Spanish would help prevent this type of crime, such as child pornography and pederasty in Mexico, leading to the arrest of these criminals.