Sentiment Analysis on Violence Against Children and Women on Social Media and News Comments in Bangladesh

Abstract

Sentiment analysis is a natural language processing task that determines the sentiment or emotion expressed in a text. In the context of violence against children and women, sentiment analysis could be used to identify comments in news articles that express negative emotions or sentiments related to positive ones. This could potentially be useful for understanding public opinion or identifying language patterns commonly used to discuss violence against children and women. Crime is now gradually expanding in Bangladesh. Various newspapers, Facebook groups, Instagram pages, and youtube videos on social media posting abuse news in post people leave comments on their expressions. Comments are expressed by their writing style and news types. This report is a research-based project on Violence against children and women sentiment analysis on social media news comments. This report will analyze people's sentiments on abusive news comments. The research will detect the sentiment of positive or negative violent news comments. Some algorithms will apply to our collected data but this paper focuses on the LSTM algorithm. In this report, the dataset will collect our own collected dataset for an expected good result. Besides the LSTM this paper contains some algorithms beside our focus. Outside LSTM there was SVM, Logistic Regression, Naive Bayes, and Naive Bayes, Random Forest was used in this report.