Jean-Baptiste/financial_news_sentiment_mixte_with_phrasebank_75

---
language:
- en
dataset_info:
splits:
- name: test
num_examples: 785
- name: train
num_examples: 4446
annotations_creators:
- expert-generated
license:
- cc-by-nc-sa-3.0
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 sentences validated by at least 75% annotators.
In addition I added ~2000 articles of Canadian news where sentiment was validated manually.

The dataset also include a column topic which contains one of the following value:
* acquisition
* other
* quaterly financial release
* appointment to new position
* dividend
* corporate update
* drillings results
* conference
* share repurchase program
* grant of stocks

This was generated automatically using a zero-shot classification model and **was not** reviewed manually.

## References
Original dataset is available here:
[https://huggingface.co/datasets/financial_phrasebank]

原始信息汇总

数据集概述

基本信息

• 名称: financial_news_sentiment_mixte_with_phrasebank_75
• 语言: 英语（English）
• 许可证: CC BY-NC-SA 3.0
• 多语言性: 单语种
• 大小类别: 1K<n<10K

任务与类别

• 任务类别: 文本分类
• 任务ID:
  • 多类分类
  • 情感分类

数据集结构

• 分割:
  • 训练集: 4446个样本
  • 测试集: 785个样本
• 注释创建者: 专家生成

数据集内容

• 附加信息: 包含一个名为“topic”的列，其值包括：
  • 收购
  • 其他
  • 季度财务发布
  • 新职位任命
  • 股息
  • 企业更新
  • 钻探结果
  • 会议
  • 股票回购计划
  • 股票授予

数据集来源

• 原始数据集: financial_phrasebank