Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CARBD-Ko: A Contextually Annotated Review Benchmark Dataset for Aspect-Level Sentiment Classification in Korean

Dongjun Jang, Jean Seo, Sungjoo Byun, Taekyoung Kim, Minseok Kim, Hyopil Shin

TL;DR

CARBD-Ko (a Contextually Annotated Review Benchmark Dataset for Aspect-Based Sentiment Classification in Korean), a benchmark dataset that incorporates aspects and dual-tagged polarities to distinguish between aspect-specific and aspect-agnostic sentiment classification, is introduced.

Abstract

This paper explores the challenges posed by aspect-based sentiment classification (ABSC) within pretrained language models (PLMs), with a particular focus on contextualization and hallucination issues. In order to tackle these challenges, we introduce CARBD-Ko (a Contextually Annotated Review Benchmark Dataset for Aspect-Based Sentiment Classification in Korean), a benchmark dataset that incorporates aspects and dual-tagged polarities to distinguish between aspect-specific and aspect-agnostic sentiment classification. The dataset consists of sentences annotated with specific aspects, aspect polarity, aspect-agnostic polarity, and the intensity of aspects. To address the issue of dual-tagged aspect polarities, we propose a novel approach employing a Siamese Network. Our experimental findings highlight the inherent difficulties in accurately predicting dual-polarities and underscore the significance of contextualized sentiment analysis models. The CARBD-Ko dataset serves as a valuable resource for future research endeavors in aspect-level sentiment classification.

CARBD-Ko: A Contextually Annotated Review Benchmark Dataset for Aspect-Level Sentiment Classification in Korean

TL;DR

CARBD-Ko (a Contextually Annotated Review Benchmark Dataset for Aspect-Based Sentiment Classification in Korean), a benchmark dataset that incorporates aspects and dual-tagged polarities to distinguish between aspect-specific and aspect-agnostic sentiment classification, is introduced.

Abstract

This paper explores the challenges posed by aspect-based sentiment classification (ABSC) within pretrained language models (PLMs), with a particular focus on contextualization and hallucination issues. In order to tackle these challenges, we introduce CARBD-Ko (a Contextually Annotated Review Benchmark Dataset for Aspect-Based Sentiment Classification in Korean), a benchmark dataset that incorporates aspects and dual-tagged polarities to distinguish between aspect-specific and aspect-agnostic sentiment classification. The dataset consists of sentences annotated with specific aspects, aspect polarity, aspect-agnostic polarity, and the intensity of aspects. To address the issue of dual-tagged aspect polarities, we propose a novel approach employing a Siamese Network. Our experimental findings highlight the inherent difficulties in accurately predicting dual-polarities and underscore the significance of contextualized sentiment analysis models. The CARBD-Ko dataset serves as a valuable resource for future research endeavors in aspect-level sentiment classification.
Paper Structure (13 sections, 4 equations, 2 figures, 4 tables)

This paper contains 13 sections, 4 equations, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. CARBD-Ko
  4. Data Collection
  5. Aspect and Opinion Extraction
  6. Manual Annotation
  7. Peer Review Stage
  8. Dataset Overview and Statistical Analysis
  9. Simultaneous Learning of Dual-Polarities via Siamese Network
  10. Experiment
  11. Settings
  12. Results
  13. Conclusion

Figures (2)

  • Figure 1: The figure provides an overview of the pipeline used to construct CARBD-Ko. It involves four steps, starting with the collection of comments from diverse domains. Next, aspect-opinion pairs are extracted from the comments. The pipeline also includes the manual annotation of both aspect-agnostic and aspect polarity, and intensity. To ensure objectivity, a peer review stage is incorporated. Overall, this pipeline enables a comprehensive sentiment analysis of the comment data in CARBD-Ko
  • Figure 2: The Operation of Siamese Network for training CARBD-Ko. While fine-tuning Transformer-based model on Sentiment Classification, the aspect-opinion token pair is passed through the Siamese Network to reduce the bias of the polarity value due to contextualization. The model is trained by simultaneously learning aspect-agnostic polarity and aspect polarity.