|
|
--- |
|
|
title: Stock Prediction Explanation |
|
|
emoji: 💬 |
|
|
colorFrom: yellow |
|
|
colorTo: purple |
|
|
sdk: gradio |
|
|
sdk_version: 5.42.0 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
hf_oauth: true |
|
|
hf_oauth_scopes: |
|
|
- inference-api |
|
|
license: apache-2.0 |
|
|
short_description: AI that predicts stock moves from news and explains why |
|
|
--- |
|
|
|
|
|
# Explainable Stock Prediction with GRPO + RAG |
|
|
|
|
|
--- |
|
|
|
|
|
## About this Space |
|
|
This demo showcases a **financial NLP model** that predicts stock market reactions to news events while also providing a **natural language explanation**. |
|
|
|
|
|
Unlike standard classifiers (which only output UP/DOWN/NEUTRAL), this model combines: |
|
|
- **GRPO (Group Relative Policy Optimization)** → reinforcement learning to refine predictions through self-reflection. |
|
|
- **RAG (Retrieval-Augmented Generation)** → retrieving past financial events to ground explanations in historical context. |
|
|
|
|
|
The result: **predictions that are both accurate and interpretable**. |
|
|
|
|
|
--- |
|
|
|
|
|
## How it works |
|
|
1. **Input** → A financial news headline or event description. |
|
|
2. **GRPO loop** → The model generates a prediction and, if wrong, reflect to improve future reasoning. |
|
|
3. **RAG retrieval** → Relevant historical events are retrieved to enrich the explanation. |
|
|
4. **Output** → |
|
|
- Chosen Stock |
|
|
- Price Prediction (UP/DOWN or percentage change) |
|
|
- Explanation (short and clear, grounded in past events) |
|
|
|
|
|
--- |
