Octa: Omissions and Conflicts in Target-Aspect Sentiment Analysis

TL;DR

Octa is a novel sentiment analysis model that jointly considers aspects and targets using a selective self-attention mechanism, effectively handling implicit or missing targets and outperforming existing models on benchmark datasets.

Contribution

It introduces a dual-layer attention approach that captures relationships between targets and context, improving sentiment inference accuracy.

Findings

Outperforms leading models by 1.6% to 4.3% in accuracy

Effectively handles implicit or missing targets

Provides a joint consideration of aspects and targets in sentiment analysis

Abstract

Sentiments in opinionated text are often determined by both aspects and target words (or targets). We observe that targets and aspects interrelate in subtle ways, often yielding conflicting sentiments. Thus, a naive aggregation of sentiments from aspects and targets treated separately, as in existing sentiment analysis models, impairs performance. We propose Octa, an approach that jointly considers aspects and targets when inferring sentiments. To capture and quantify relationships between targets and context words, Octa uses a selective self-attention mechanism that handles implicit or missing targets. Specifically, Octa involves two layers of attention mechanisms for, respectively, selective attention between targets and context words and attention over words based on aspects. On benchmark datasets, Octa outperforms leading models by a large margin, yielding (absolute) gains in accuracy of 1.6% to 4.3%.