Helping users discover perspectives: Enhancing opinion mining with joint topic models

TL;DR

This paper investigates how joint topic models can improve opinion mining by identifying distinct perspectives in contentious issues, demonstrating that models like TAM align well with human judgments and are unaffected by users' pre-existing biases.

Contribution

It evaluates four joint topic models for perspective discovery, showing TAM effectively captures perspectives consistent with human understanding in opinion mining.

Findings

TAM aligns with human judgments of perspectives.

Users' pre-existing biases do not influence interpretation of model output.

Joint topic models can effectively identify perspectives in opinionated texts.

Abstract

Support or opposition concerning a debated claim such as abortion should be legal can have different underlying reasons, which we call perspectives. This paper explores how opinion mining can be enhanced with joint topic modeling, to identify distinct perspectives within the topic, providing an informative overview from unstructured text. We evaluate four joint topic models (TAM, JST, VODUM, and LAM) in a user study assessing human understandability of the extracted perspectives. Based on the results, we conclude that joint topic models such as TAM can discover perspectives that align with human judgments. Moreover, our results suggest that users are not influenced by their pre-existing stance on the topic of abortion when interpreting the output of topic models.