Answers are grounded in this paper and cite their sources.
Demonstrates enhanced hate speech detection using a novel hybrid model, indicating significant accuracy improvement.
Hate speech is a fundamental societal obstacle that creates emotional and psychological impacts on public policies and social harmony. Interaction with different groups and individuals may generate hate and racist text, which influences inter-group threat and the degree of anxiety. Hate speech-dependent features with contextual analysis and level of anxiety in communication with threat propagation are overlooked presently. The novel research proposed a hybrid framework named HateAnxITT to overcome these challenges with a novel hate and racism dependent feature with two linguistic layers of inter-group anxiety and threat theory, paragraph embedding model, and Fuzzy rough nearest neighbour (FRNN) classifier considering OWA operator. This research proposed a novel feature to assess (i) inter-group anxiety feature with tone analyser, dominant attitudes and negative emotions, and (ii) inter-group and individual threat linguistic feature with POS tagger, typed dependency and both-sided pronouns. Along with these two features, the paragraph embedding model is considered for the contextual analysis of text. The proposed HateAnxITT model considers input features and values with the FRNN-OWA classifier. The performance analysis considered present state-of-the-art models and existing NLP features, where the experimentation and evaluation section shows that the proposed model improved up to 7% accuracy on the standard dataset for hate and racism.
No takes yet. Share an insight, caveat, or question.
Ghosal et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: