README.md

metadata

annotations_creators: []
language:
  - en
language_creators: []
license: []
multilinguality:
  - monolingual
pretty_name: adaption-MITRA_FinReg
size_categories:
  - 10K<n<100K
source_datasets:
  - original
tags:
  - adaption
  - instruction-tuning
  - corporate-business
  - market-analysis
  - news
task_categories: []
task_ids: []

This dataset is a remastered version prepared using Adaption's Adaptive Data platform.

adaption-MITRA_FinReg

This dataset consists of short news snippets and corporate announcements paired with sentiment labels classified as positive, negative, or neutral. The content covers diverse business topics including service descriptions, financial growth, legal disputes, and technology deployments. It is formatted as prompt-completion pairs suitable for training or evaluating sentiment analysis models in a financial or corporate context.

Dataset size

There are 28,298 data points in this dataset. This is an instruction tuning dataset.

Quality of Remastered Dataset

The final quality is B, with a relative quality improvement of 170.0%.

Domain

• Corporate-business (42%)
• Market-analysis (16%)
• News (16%)

Language

• English (100%)

Tone

• Informative (79%)
• Professional (11%)
• Urgent (5%)

Evaluation Results

• Quality Gains:

  

• Grade Improvement:

  

• Percentile Chart: