Datasets:

Samsoup
/

yelp2018

dataset_info:
  features:
    - name: text
      dtype: string
    - name: label
      dtype: float32
  splits:
    - name: train
      num_examples: 7000
    - name: validation
      num_examples: 1500
    - name: test
      num_examples: 1500
  task_categories:
    - text-regression
  task_ids:
    - sentiment-regression

Samsoup/yelp2018 (Regression)

A sentiment regression dataset derived from the Kaggle Yelp Reviews Dataset (https://www.kaggle.com/datasets/omkarsabnis/yelp-reviews-dataset), split into 70% train / 15% validation / 15% test.

Fields

text – review text
label – float in [1.0, 5.0] (the original star rating)

Preparation We kept only text and stars, renamed the target to label (kept in 1..5 as float), and created a stratified 70/15/15 split by the discrete star bins (1..5).

Citation Yuxia Wang, Daniel Beck, Timothy Baldwin, and Karin Verspoor. 2022. Uncertainty Estimation and Reduction of Pre-trained Models for Text Regression. TACL 10:680–696. https://doi.org/10.1162/tacl_a_00492

Source Kaggle – Yelp Reviews Dataset: https://www.kaggle.com/datasets/omkarsabnis/yelp-reviews-dataset

Last updated: 2025-10-15