# market-sentiment-roberta-pro

## Overview
Market-Sentiment-RoBERTa-Pro is a high-precision sentiment classifier for financial news, social media posts, and earnings call transcripts. It distinguishes between Bullish, Neutral, and Bearish market sentiments.

## Model Architecture
- **Base Model:** `roberta-base`
- **Data:** Trained on the Financial PhraseBank and 5M+ scraped financial tweets.
- **Optimization:** AdamW optimizer with a linear learning rate decay.

## Intended Use
- Real-time algorithmic trading signals.
- Sentiment monitoring for hedge funds and retail investors.
- Correlation analysis between social sentiment and price action.

## Limitations
- May misinterpret high-frequency trading sarcasm.
- Context window is limited to 512 tokens; very long earnings reports should be chunked.

## Example Code
```python
from transformers import pipeline

sentiment_analyser = pipeline("sentiment-analysis", model="market-sentiment-roberta-pro")
tweet = "The quarterly revenue for $AAPL exceeded expectations, suggesting a strong growth trajectory."
print(sentiment_analyser(tweet))