metadata

dataset_info:
  features:
    - name: comment
      dtype: string
    - name: quad
      sequence:
        sequence: string
    - name: dataset
      dtype: string
  splits:
    - name: train
      num_bytes: 2111953
      num_examples: 3987
    - name: test
      num_bytes: 266209
      num_examples: 500
    - name: validation
      num_bytes: 88525
      num_examples: 170
  download_size: 1136999
  dataset_size: 2466687
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
      - split: validation
        path: data/validation-*

OATS Dataset

Description

The OATS (Opinion Aspect Target Sentiment) dataset is a comprehensive collection designed for the Aspect Sentiment Quad Prediction (ASQP) or Aspect-Category-Opinion-Sentiment (ACOS) task. This dataset aims to facilitate research in aspect-based sentiment analysis by providing detailed opinion quadruples extracted from review texts. Additionally, for each review, we offer tuples summarizing the dominant sentiment polarity toward each aspect category discussed.

The dataset covers three distinct domains: Amazon FineFood reviews, Coursera course reviews, and TripAdvisor Hotel reviews, offering a broad spectrum for analysis across different types of services and products. Structure

The dataset is structured into two primary components:

Opinion Quadruples: Detailed annotations on the level of individual opinions, including the aspect, the sentiment target, and the corresponding sentiment.
Review-Level Tuples: Aggregate information at the review level, indicating the overall sentiment polarity for each aspect category mentioned.

Domains

Amazon FineFood Reviews
Coursera Course Reviews
TripAdvisor Hotel Reviews

Each domain is annotated from scratch, ensuring high-quality data for nuanced sentiment analysis tasks. Citation

If you use the OATS dataset in your research, please cite the original authors:

@misc{chebolu2023oats,
    title={OATS: Opinion Aspect Target Sentiment Quadruple Extraction Dataset for Aspect-Based Sentiment Analysis},
    author={Siva Uday Sampreeth Chebolu and Franck Dernoncourt and Nedim Lipka and Thamar Solorio},
    year={2023},
    eprint={2309.13297},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

Usage

This dataset has been curated to facilitate easy access and integration into existing NLP pipelines. To use this dataset, you can load it using the datasets library by Hugging Face:

from datasets import load_dataset

dataset = load_dataset("jordiclive/OATS-ABSA")