metadata

language:
  - en
dataset_info:
  splits:
    - name: test
      num_examples: 785
    - name: train
      num_examples: 4446
annotations_creators:
  - expert-generated
license:
  - mit
multilinguality:
  - monolingual
pretty_name: financial_news_sentiment_mixte_with_phrasebank_75
size_categories:
  - 1K<n<10K
tags: []
task_categories:
  - text-classification
task_ids:
  - multi-class-classification
  - sentiment-classification

Dataset Card for "financial_news_sentiment_mixte_with_phrasebank_75"

This is a customized version of the phrasebank dataset in which I kept only sentence validated by at least 75% annotators.
In addition I added ~2000 articles validated manually on Canadian financial news.