| --- |
| language: |
| - kk |
| license: apache-2.0 |
| tasks: |
| - text-classification |
| multilinguality: |
| - monolingual |
| size_categories: |
| - 10K<n<100K |
| task_categories: |
| - text-classification |
| --- |
| |
| # Darmm Sentiment KK |
|
|
| Kazakh sentiment dataset with 5 labels for multi-domain reviews. |
|
|
| ## Labels |
|
|
| - `very_negative` |
| - `negative` |
| - `neutral` |
| - `positive` |
| - `very_positive` |
|
|
| ## Data fields |
|
|
| - `id`: unique identifier |
| - `text`: Kazakh text |
| - `label`: sentiment label |
| - `source`: data source reference |
| - `domain`: domain tag (entertainment, apps, services, commerce, etc.) |
|
|
| ## Splits |
|
|
| - `train`, `validation`, `test` |
|
|
| ## Curation |
|
|
| Texts are collected and labeled manually with consistency guidelines. |
