el-oc-tweets

This notebook presents a solution for the SemEval-2018 Task 1: Affect in Tweets (AIT-2018). Given a tweet and an emotion (sadness, anger, fear, joy), the model should output an intensity classification (0, 1, 2, 3). To do this, two different approaches were implemented: the first one using a knowledge based preprocessing the second one using word embeddings (BERT). Finally, several models were used and compared to do the classification.