Views
No views yet
xlm-roberta-base on a self-collected dataset of product reviews scraped from Uzum Market (uzbekistan's largest e-commerce platform).| Class | Precision | Recall | F1-Score | Support |
|---|---|---|---|---|
| Negative | 0.7586 | 0.8800 | 0.8148 | 25 |
| Positive | 0.9663 | 0.9247 | 0.9451 | 93 |
| Weighted avg | 0.9223 | 0.9153 | 0.9175 | 118 |
1from transformers import pipeline
2
3classifier = pipeline(
4 "text-classification",
5 model="sant1x/uzbek-sentiment-xlm-roberta"
6)
7
8classifier("Mahsulot juda yaxshi, sifati a'lo!")
9# [{'label': 'pos', 'score': 0.97}]Norboev, S. (2025). Uzbek Sentiment Analysis using XLM-RoBERTa.
Hugging Face. https://huggingface.co/sant1x/uzbek-sentiment-xlm-roberta