mohanrj/mybully-sentibert-manual

MYBully-SentiBERT (Manual)

Model Overview

This model is MYBully-SentiBERT, fine-tuned exclusively on the manually annotated portion of the MYBully dataset.
It provides a baseline for Sentiment Analysis in Bahasa Malaysia tweets.

Intended Use

• Sentiment Classification (positive, negative, neutral).

Training Data

• Dataset: MYBully (Bahasa Malaysia tweets).
• Annotation: Manual

Model Details

• Base model: roberta-base-bahasa-cased
• Fine-tuning: 3-way classification head
• Labels: Positive, Negative, Neutral

Performance

Metric	Value
Accuracy	0.83
Precision	0.82
Recall	0.83
F1	0.82