---
license: mit
tags:
  - nifty
  - stock-movement
  - news-and-events
  - RLMF
task_categories:
  - multiple-choice
  - time-series-forecasting
  - document-question-answering
task_ids:
  - topic-classification
  - semantic-similarity-classification
  - multiple-choice-qa
  - univariate-time-series-forecasting
  - document-question-answering
language:
  - en
pretty_name: nifty-rl
size_categories:
  - 1K<n<100k
configs:
- config_name: nifty-rl
  data_files:
    - split: train
      path: "train.jsonl"
    - split: test
      path: "test.jsonl"
    - split: valid
      path: "valid.jsonl"
  default: true
  
---

<h1>
  <img alt="RH" src="./nifty-icon.png" style="display:inline-block; vertical-align:middle; width:120px; height:120px; object-fit:contain" />
    The News-Informed Financial Trend Yield (NIFTY) Dataset. 
</h1>

The News-Informed Financial Trend Yield (NIFTY) Dataset. Details of the dataset, including data procurement and filtering can be found in the paper here: https://arxiv.org/abs/2405.09747.


## 📋 Table of Contents

- [🧩 NIFTY Dataset](#nifty-dataset)
  - [📋 Table of Contents](#table-of-contents)
  - [📖 Usage](#usage)
    - [Downloading the dataset](#downloading-the-dataset)
    - [Dataset structure](#dataset-structure)
  - [Large Language Models](#large-language-models) 
  - [✍️ Contributing](#contributing)
  - [📝 Citing](#citing)
  - [🙏 Acknowledgements](#acknowledgements)

## 📖 [Usage](#usage)

Downloading and using this dataset should be straight-forward following the Huggingface datasets framework. 

### [Downloading the dataset](#downloading-the-dataset)

The NIFTY dataset is available on huggingface [here](https://huggingface.co/datasets/raeidsaqur/NIFTY) and can be downloaded with the following python snipped:

```python

from datasets import load_dataset

# If the dataset is gated/private, make sure you have run huggingface-cli login
dataset = load_dataset("raeidsaqur/nifty-rl")

```

### [Dataset structure](#dataset-structure)

The dataset is split into 3 partition, train, valid and test and each partition is a jsonl file where a single row has the following keys.

```python
['prompt', 'chosen', 'rejected', 'chosen_label', 'chosen_value']
```

Currently, the dataset has 2111 examples in total, the dates randing from 2010-01-06 to 2020-09-21. 
<!-- The number of examples for each split is given below.
| Split | Num Examples | Date range |
|-------|--------------|------------|
|Train |1477 |2010-01-06 - 2017-06-27 |
|Valid|317 | 2017-06-28- 2019-02-12|
|Test |317|2019-02-13 - 2020-09-21|
 -->
<!--
<img alt="St" src="./imgs/visualize_nifty_1794_2019-02-13.png" 
  style="display:inline-block; vertical-align:middle; width:640px; 
  height:640px; object-fit:contain" />

-->

 

## ✍️  [Contributing](#contributing)

We welcome contributions to this repository (noticed a typo? a bug?). To propose a change:

```
git clone https://huggingface.co/datasets/raeidsaqur/nifty-rl
cd nifty-rl
git checkout -b my-branch
pip install -r requirements.txt
pip install -e .
```

Once your changes are made, make sure to lint and format the code (addressing any warnings or errors):

```
isort .
black .
flake8 .
``` 

Then, submit your change as a pull request. 

## 📝  [Citing](#citing)

If you use the NIFTY Financial dataset in your work, please consider citing our paper:

```
@article{raeidsaqur2024NiftyRL,
    title        = {NIFTY-RL: Financial News Headlines Dataset for LLM Alignment using Reinforcement Learning.},
    author       = {Raeid Saqur},
    year         = 2024,
    journal      = {ArXiv},
    url          = {https://arxiv.org/abs/2024.5599314}
}
```

## 🙏 [Acknowledgements](#acknowledgements)

The authors acknowledge and thank the generous computing provided by the Vector Institute, Toronto.