Socio-analyzer: A sentiment analysis of #MeToo tweets using artificial recurrent neural network

Abstract

Social media has developed into a highly effective tool for exchanging information. They enable users to share and discuss a variety of topics of interest in addition to consuming information. On the other hand, social media has some drawbacks, such as trolling and online threats. Twitter is a social networking site where users exchange short messages known as tweets. There are numerous applications for tweets. Trolls and bullies post disparaging comments on social media to annoy, cause trouble, or actively harm others. #MeToo has recently become one of Twitter's most popular trending topics, with many tweets being tagged with the hashtag. The goal of the movement is to raise public awareness about sexual assault and harassment, empower women to report incidents of sexual harassment and abuse and provide assistance to those who have been victimized. Thirty thousand tweets were used in our study, and we divided them into two categories: hateful and not. Different models, such as Naive Bayes, Logistic Regression, and LSTM, were used to collect data for sentiment classification. With LSTM we achieved an accuracy of 97.90 % in classifying hateful and non-hateful tweets. These methods make the detection of troll and hateful MeToo tweets easier. These technological innovations aid in retrieving information from tweets and preventing harmful propaganda against the MeToo movement, discouraging women from speaking out against sexual abusers.