Sentiment Analysis of Public Perception Toward Mob Justice

Mob justice, characterized by groups bypassing legal systems to enforce immediate punishment, undermines the rule of law and human rights. This study leverages natural language processing (NLP) to classify public sentiment on mob justice using a dataset of 42,277 annotated Facebook posts and comments. Sentiment is labeled as Positive, Negative, and Neutral. We applied several traditional machine learning i.e. Logistic Regression, SVM, deep learning model BiLSTM, and transformer-based model BERT, RoBERTa, DistilBERT, ELECTRA. DistilBERT outperformed others with an F1-score of 0.93, demonstrating how well contextual embeddings capture complex emotions. The analysis reveals principal sentiment trends, offering insights into public perceptions that can guide policy and social discourse on this crucial socio-legal issue.