Identification of Misogyny on Social Media in Indonesian Using Bidirectional Encoder Representations From Transformers (BERT)

Abstract

Misogyny is a behavior that hates or dislikes women Text classification can be used to identify misogyny text. One text classification method currently popular and proven to have good performance is the Bidirectional Encoder From Transformers (BERT). Fine-tuning is a method to transfer knowledge from a trained model to a new model to complete a new task. This study focuses on building a misogyny identification model with IndoBert pre-trained model provided by IndoNLU. The identification of Misogyny model obtained the best results with an accuracy value of 83.74% and by using K-fold cross-validation, the average validation value is 77.86%.