Opinion Mining Using Population-tuned Generative Language Models

TL;DR

This paper introduces a new opinion mining approach leveraging population-tuned generative language models, which can learn and transfer opinions across semantic classes while preserving polarization, demonstrated through experiments and real text corpus analysis.

Contribution

The paper proposes a novel method for opinion insight mining using fine-tuned generative language models trained on population-specific data, enhancing opinion transfer and scalability.

Findings

Effective transfer of opinions to semantic classes

Maintains polarization proportions in opinion transfer

Scalable opinion insight discovery from large text corpora

Abstract

We present a novel method for mining opinions from text collections using generative language models trained on data collected from different populations. We describe the basic definitions, methodology and a generic algorithm for opinion insight mining. We demonstrate the performance of our method in an experiment where a pre-trained generative model is fine-tuned using specifically tailored content with unnatural and fully annotated opinions. We show that our approach can learn and transfer the opinions to the semantic classes while maintaining the proportion of polarisation. Finally, we demonstrate the usage of an insight mining system to scale up the discovery of opinion insights from a real text corpus.