Leveraging Structural and Semantic Correspondence for Attribute-Oriented Aspect Sentiment Discovery

TL;DR

This paper introduces Trait, an unsupervised probabilistic model that leverages structural and semantic correspondence to improve aspect and sentiment discovery in opinionated text, explicitly incorporating attributes for more accurate and coherent results.

Contribution

The paper presents Trait, a novel unsupervised model that explicitly incorporates attribute-related structural and semantic correspondence using a Markov Random Field, enhancing aspect and sentiment discovery.

Findings

Trait outperforms state-of-the-art baselines in attribute profile accuracy

Trait generates more semantically cohesive topics

Visualization confirms the alignment of attribute profiles with intuitions

Abstract

Opinionated text often involves attributes such as authorship and location that influence the sentiments expressed for different aspects. We posit that structural and semantic correspondence is both prevalent in opinionated text, especially when associated with attributes, and crucial in accurately revealing its latent aspect and sentiment structure. However, it is not recognized by existing approaches. We propose Trait, an unsupervised probabilistic model that discovers aspects and sentiments from text and associates them with different attributes. To this end, Trait infers and leverages structural and semantic correspondence using a Markov Random Field. We show empirically that by incorporating attributes explicitly Trait significantly outperforms state-of-the-art baselines both by generating attribute profiles that accord with our intuitions, as shown via visualization, and yielding topics of greater semantic cohesion.