From Sentiment Annotations to Sentiment Prediction through Discourse Augmentation

TL;DR

This paper introduces a novel discourse-augmented neural framework for sentiment analysis of long documents, improving performance over existing methods by leveraging discourse structures and ensemble techniques.

Contribution

It presents a new neural architecture combining discourse augmentation with a TreeLSTM model for better sentiment prediction in long texts.

Findings

Enhanced sentiment prediction accuracy for long documents.

Discourse augmentation outperforms traditional discourse parsers.

Ensemble methods further improve performance based on document length.

Abstract

Sentiment analysis, especially for long documents, plausibly requires methods capturing complex linguistics structures. To accommodate this, we propose a novel framework to exploit task-related discourse for the task of sentiment analysis. More specifically, we are combining the large-scale, sentiment-dependent MEGA-DT treebank with a novel neural architecture for sentiment prediction, based on a hybrid TreeLSTM hierarchical attention model. Experiments show that our framework using sentiment-related discourse augmentations for sentiment prediction enhances the overall performance for long documents, even beyond previous approaches using well-established discourse parsers trained on human annotated data. We show that a simple ensemble approach can further enhance performance by selectively using discourse, depending on the document length.