The objective of a multi-stage machine learning and fuzzy approach to cyber-hate detection is to develop a robust and effective system for identifying and mitigating instances of cyber-hate speech or harmful content on digital platforms.
FRAMEWORK:
DATASETS USED
FORM SPRING DATASET Consists of both hate and non-hate texts in a tabulated manner
MARYLAND DATASET Main recognition of linguistic features and to study nature and culture abusive comments.
DAVIDSON DATASET Worst possible data consists of racism, misogynistic, violent and homophobic.
OLID DATASET Consists of English tweets consists of 14,100 tweets!