Сентиментальний аналіз як інструмент для виявлення шкідливого контенту

Abstract

This work is devoted to sentiment analysis methods to detect malicious content in personal communications like social networks, e-mail and instant messengers. The features of manipulative techniques in phishing messages, fake pranks and spam messages are considered. A model is proposed that combines the analysis of emotional coloring with the identification of linguistic patterns characteristic of malicious content of a personal nature. Also, it shows where the research results can be applied.