Seernet at EmoInt-2017: Tweet Emotion Intensity Estimator

TL;DR

This paper presents a system for estimating emotion intensity in tweets by combining lexical, syntactic, and embedding features, achieving third place in a shared task competition.

Contribution

It introduces an ensemble approach that integrates multiple feature types and regressors for improved emotion intensity estimation in social media text.

Findings

Achieved third place in WASSA-2017 Shared Task

Effective combination of lexical, syntactic, and embedding features

Ensemble method outperforms individual models

Abstract

The paper describes experiments on estimating emotion intensity in tweets using a generalized regressor system. The system combines lexical, syntactic and pre-trained word embedding features, trains them on general regressors and finally combines the best performing models to create an ensemble. The proposed system stood 3rd out of 22 systems in the leaderboard of WASSA-2017 Shared Task on Emotion Intensity.