Dartmouth CS at WNUT-2020 Task 2: Informative COVID-19 Tweet Classification Using BERT
share
We describe the systems developed for the WNUT-2020 shared task 2, identification of informative COVID-19 English Tweets. BERT is a highly performant model for Natural Language Processing tasks. We increased BERT's performance in this classification task by fine-tuning BERT and concatenating its embeddings with Tweet-specific features and training a Support Vector Machine (SVM) for classification (henceforth called BERT+). We compared its performance to a suite of machine learning models. We used a Twitter specific data cleaning pipeline and word-level TF-IDF to extract features for the non-BERT models. BERT+ was the top performing model with an F1-score of 0.8713.
READ FULL TEXT