README.md

---
dataset_info:
  features:
  - name: ID
    dtype: string
  - name: Text
    dtype: string
  - name: Polarity
    dtype: string
  - name: Domain
    dtype: string
  splits:
  - name: train
    num_bytes: 15164685
    num_examples: 70000
  download_size: 7415117
  dataset_size: 15164685
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
tags:
- sentigold
- bangla
- bangladataset
- sentiment
---


# Bangla Sentiment Dataset (SentiGOLD v1)

**SentiGOLD 70k v1** A multi-domain sentiment analysis dataset in Bangla is called SentiGOLD.  A gender-balanced team of linguists annotated the 70,000 samples, which were obtained from a variety of sources.  SentiGOLD complies with language standards that have been developed by a Bangla linguistics commission and the Government of Bangladesh.

Each text entry in the dataset is categorized into one of the following sentiment classes:

- **SP**: Strongly Positive 
- **WP**: Weakly Positive
- **WN**: Weakly Positive Negative  
- **SN**: Strongly Negative  
- **NU**: Neutral  

This dataset provides a valuable resource for building and evaluating sentiment analysis models in the Bangla language.


## Use the Dataset

```python
from datasets import load_dataset

dataset = load_dataset('SayedShaun/sentigold')
print(dataset)

>>> DatasetDict({
>>>     train: Dataset({
>>>         features: ['ID', 'Text', 'Polarity', 'Domain'],
>>>         num_rows: 70000
>>>    })
>>> })
```


## Source and Citation

**[SentiGOLD: A Large Bangla Gold Standard Multi-Domain Sentiment Analysis Dataset and Its Evaluation](https://arxiv.org/abs/2306.06147)**  

```bibtex
@inproceedings{islam2023sentigold,
  title={Sentigold: A large bangla gold standard multi-domain sentiment analysis dataset and its evaluation},
  author={Islam, Md Ekramul and Chowdhury, Labib and Khan, Faisal Ahamed and Hossain, Shazzad and Hossain, Md Sourave and Rashid, Mohammad Mamun Or and Mohammed, Nabeel and Amin, Mohammad Ruhul},
  booktitle={Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
  pages={4207--4218},
  year={2023}
}