Data Utility Evaluation of Different Misogyny Datasets Using Machine Learning Models and Extracting Feature Approaches

Social networks are highly used for social interactions. Monitoring and verifying the content of data that spreads over them can contribute to limiting the spread and fomenting hate speech, which is a phenomenon that requires the development of new methodologies and strategies to discriminate against hate speech in textual content. To this end, text data, especially from social network platforms, can benefit from analytical activities devoted to hate speech discrimination, such as misogyny. In this proposal, we organized and cleaned misogyny datasets collected from online sources and provided classification results obtained by employing machine learning models and feature extraction approaches. Experimental results produced by employing several machine learning models on the misogyny datasets reveal discrimination improvements when feature extraction approaches are used. The proposal aims to support stakeholders, data analysts, and researchers by providing them with clean misogyny datasets, organized for analytical activities, together with statistical classification results obtained using machine learning models and feature extraction approaches.