LT3 at SemEval-2018 Task 1: A classifier chain to detect emotions in tweets

Abstract

This paper presents an emotion classification system for English tweets, submitted for the SemEval shared task on Affect in Tweets, sub-task 5: Detecting Emotions. The system combines lexicon, n-gram, style, syntactic and semantic features. For this multi-class multilabel problem, we created a classifier chain. This is an ensemble of eleven binary classifiers, one for each possible emotion category, where each model gets the predictions of the preceding models as additional features. The predicted labels are combined to get a multilabel representation of the predictions. Our system was ranked eleventh among thirty five participating teams, with a Jaccard accuracy of 52.0% and macro- and micro-average F1-scores of 49.3% and 64.0%, respectively.