README.md

---
task_categories:
- text-classification
language:
- vi
---
## Dataset Card for UIT‑VSMEC

### 1. Dataset Summary

**UIT‑VSMEC** (Vietnamese Social Media Emotion Corpus) is a benchmark corpus for emotion recognition in Vietnamese social media comments. It consists of **6,927** human‐annotated sentences, each labeled with one of six basic emotions, plus an `Other` category.

### 2. Supported Tasks and Leaderboard

* **Primary Task**: Text classification – emotion recognition
* **Metrics**: Accuracy, F1‑score

*No public leaderboard yet; contributions welcome!*

### 3. Languages

* Vietnamese

### 4. Data Fields

* **Sentence** (`str`): The raw social media comment.
* **Emotion** (`str`): Emotion category.
* **type** (`str`): Split name.
* **index** (`int`): The index of sentence.

### 5. Emotion Labels

| Label       | Description                    |
| ----------- | ------------------------------ |
| `Enjoyment` | Joy, happiness, or pleasure.   |
| `Sadness`   | Grief, sorrow, or unhappiness. |
| `Anger`     | Annoyance or hostility.        |
| `Fear`      | Anxiety or apprehension.       |
| `Disgust`   | Revulsion or aversion.         |
| `Surprise`  | Shock or amazement.            |
| `Other`     | None of the above.             |


### 6. Usage

```python
from datasets import load_dataset

ds = load_dataset("visolex/UIT-VSMEC")

train_ds = ds.filter(lambda x: x["type"] == "train")
dev_ds   = ds.filter(lambda x: x["type"] == "dev")
test_ds  = ds.filter(lambda x: x["type"] == "test")

print(train_ds[0])
```

### 7. Dataset Creation

1. **Source Data**: Original splits were downloaded from a Google Drive folder.
2. **Merge Process**: Combined train/dev/test CSVs into one file; added a `type` column.
3. **Preprocessing**: Minimal cleaning—only normalized whitespace; preserved original labels.

### 8. Source & Links

* **Original Google Drive** (raw CSVs):
  [https://drive.google.com/drive/folders/1HooABJyrddVGzll7fgkJ6VzkG\_XuWfRu](https://drive.google.com/drive/folders/1HooABJyrddVGzll7fgkJ6VzkG_XuWfRu)

### 9. Licenses and Citation

#### License

Please refer to the original dataset license (if unspecified, assume **CC BY 4.0**).

#### How to Cite

**Original Paper**

```
@InProceedings{10.1007/978-981-15-6168-9_27,
  author    = {Ho, Vong Anh and Nguyen, Duong Huynh-Cong and Nguyen, Danh Hoang and Pham, Linh Thi-Van
               and Nguyen, Duc-Vu and Nguyen, Kiet Van and Nguyen, Ngan Luu-Thuy},
  title     = {Emotion Recognition for Vietnamese Social Media Text},
  booktitle = {Computational Linguistics},
  year      = {2020},
  publisher = {Springer Singapore},
  pages     = {319--333},
  isbn      = {978-981-15-6168-9}
}
```