SentiME++ at SemEval-2017 Task 4A: Stacking state-of-the-art classifiers to enhance sentiment classification

Palumbo, Enrico; Sygkounas, Efstratios; Troncy, Raphaël; Rizzo, Giuseppe
SEMEVAL 2017, 11th International Workshop on Semantic Evaluation, collocated with the 55th annual meeting of the Association for Computational Linguistics (ACL), August 3-4, 2017, Vancouver, Canada

In this paper, we describe the participation of the SentiME++ system to the SemEval
2017 Task 4A "Sentiment Analysis in Twitter" that aims to classify whether
English tweets are of positive, neutral or negative sentiment. SentiME++ is an ensemble
approach to sentiment analysis that leverages stacked generalization to automatically
combine the predictions of five state-of-the-art sentiment classifiers. SentiME++
achieved officially 61.30% F1-score, ranking 12th out of 38 participants.

Type:
Conference
City:
Vancouver
Date:
2017-08-03
Department:
Data Science
Eurecom Ref:
5237
Copyright:
Copyright ACL. Personal use of this material is permitted. The definitive version of this paper was published in SEMEVAL 2017, 11th International Workshop on Semantic Evaluation, collocated with the 55th annual meeting of the Association for Computational Linguistics (ACL), August 3-4, 2017, Vancouver, Canada and is available at :

PERMALINK : https://www.eurecom.fr/publication/5237