Classification

Introduction

The data was modelled in two different approaches to test multiple classifiers. First a bag of words modelling and secondly through different BERT transformers implementation.

Classification

This work will experiment on:

• Training a supervised classifier with a labeled set of tweets to predict if a tweet is misogynistic or not. Test the classifier with a set of tweets not used for training and test it with a small set of new articles to evaluate its precision. The objective is to find out if training models with tweets can be useful for different text entries.

• In terms of models, as a baseline a NB model using Bag of Words modelling will be implemented to compare its results with the ones obtained through fine-tuning different pretrained BERT models.