---
license: cc-by-nc-sa-3.0
task_categories:
- text-classification
language:
- mk
multilinguality:
- monolingual
pretty_name: Macedonian Tweet Sentiment Classification
size_categories:
- 1K<n<10K
---

# Dataset Card for Macedonian Tweet Sentiment Classification

## Dataset Description

### Dataset Summary
This is a Macedonian dataset is a collection of tweets for sentiment classification. 

### Supported Tasks and Leaderboards
text-classification, sentiment-classification: The dataset can be used to train a model for sentiment classification. The model performance is evaluated based on the accuracy of the predicted labels as compared to the given labels in the dataset. The BCP-47 code for Macedonian is mk. 

### Languages
The text is in Macedonian.

## Dataset Structure

### Data Fields
- `text`: String of a tweet.
- `label`: An integer indicating annotated sentiment level: -1 for negative, 0 for neutral, and 1 for positive.

### Data Splits

The dataset has 2 splits: _train_, and _test_. 

| Dataset | Positive | Neutral | Negative | Total |
|---------|----------|---------|----------|-------|
| Train   | 2,610 (30%) | 1,280 (15%) | 4,693 (55%) | 8,583 |
| Test    | 431 (38%) | 200 (18%) | 508 (44%) | 1,139 |


## Dataset Creation
Please see the [paper](https://aclanthology.org/R15-1034/) for details.

## Additional Information

### Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
```
@inproceedings{jovanoski-etal-2015-sentiment,
    title = "Sentiment Analysis in {T}witter for {M}acedonian",
    author = "Jovanoski, Dame  and
      Pachovski, Veno  and
      Nakov, Preslav",
    editor = "Mitkov, Ruslan  and
      Angelova, Galia  and
      Bontcheva, Kalina",
    booktitle = "Proceedings of the International Conference Recent Advances in Natural Language Processing",
    month = sep,
    year = "2015",
    address = "Hissar, Bulgaria",
    publisher = "INCOMA Ltd. Shoumen, BULGARIA",
    url = "https://aclanthology.org/R15-1034",
    pages = "249--257",
}
```