OFFENSIVE LANGUAGE DETECTION THROUGH TEXT CLASSIFICATION

In this book, we propose a modular text classification pipeline on social media datasets focusing on Twitter. Our proposed approach is to leverage a modular development that allows easy use for combining different text classification components. This work's main contribution is that it presents a new modular text classification pipeline to facilitate benchmarking by conducting a detailed analytical study of the best-performing approaches, features, and embeddings reported by the state-of-the-art.