fancyzhx/amazon_polarity

Dataset Overview

Dataset Name

• Name: Amazon Review Polarity
• Alias: AmazonPolarity

Basic Information

• Language: English (en)
• License: Apache‑2.0
• Multilinguality: Monolingual
• Size: 1 M < n < 10 M
• Source Data: Raw data
• Task Category: Text classification
• Task ID: Sentiment classification

Dataset Structure

• Features:
  • label: Classification label, 0 = negative, 1 = positive
  • title: String, review title
  • content: String, review content
• Splits:
  • Training set: 3,600,000 samples, total size 1,604,364,432 bytes
  • Test set: 400,000 samples, total size 178,176,193 bytes

Usage

• Training / Evaluation Metrics:
  • Accuracy
  • F1 (macro, micro, weighted)
  • Precision (macro, micro, weighted)
  • Recall (macro, micro, weighted)

Creation

• Creator: Xiang Zhang (xiang.zhang@nyu.edu)
• Purpose: Benchmark for text classification, used in the paper "Character‑level Convolutional Networks for Text Classification"

License Information

• License: Apache License 2.0

Citation

• McAuley, Julian, and Jure Leskovec. "Hidden factors and hidden topics: understanding rating dimensions with review text." In Proceedings of the 7th ACM conference on Recommender systems, pp. 165‑172. 2013.
• Xiang Zhang, Junbo Zhao, Yann LeCun. "Character‑level Convolutional Networks for Text Classification." Advances in Neural Information Processing Systems 28 (NIPS 2015).