A Multi-task Ensemble Framework for Emotion, Sentiment and Intensity Prediction

TL;DR

This paper introduces a multi-task ensemble framework combining deep learning models and handcrafted features to improve emotion and sentiment analysis across various tasks, domains, and granularities.

Contribution

It presents a novel multi-task ensemble approach that leverages CNN, LSTM, GRU, and handcrafted features for enhanced emotion and sentiment prediction.

Findings

Achieved 2-3 point performance improvements over single-task models

Effective across diverse domains like social media and news

Demonstrated robustness on benchmark datasets

Abstract

In this paper, through multi-task ensemble framework we address three problems of emotion and sentiment analysis i.e. "emotion classification & intensity", "valence, arousal & dominance for emotion" and "valence & arousal} for sentiment". The underlying problems cover two granularities (i.e. coarse-grained and fine-grained) and a diverse range of domains (i.e. tweets, Facebook posts, news headlines, blogs, letters etc.). The ensemble model aims to leverage the learned representations of three deep learning models (i.e. CNN, LSTM and GRU) and a hand-crafted feature representation for the predictions. Experimental results on the benchmark datasets show the efficacy of our proposed multi-task ensemble frameworks. We obtain the performance improvement of 2-3 points on an average over single-task systems for most of the problems and domains.