A reinforcement-learning trading agent that fuses numerical market signals with natural-language news sentiment, deployed via FastAPI with full model interpretability.
Portfolio decisions are sequential and uncertain. Real systems must integrate structured time-series data with unstructured information such as news.
This project implements a reinforcement-learning agent that learns when to invest or stay in cash using return regimes and sentiment polarity.
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
pip install python-multipart
python train.py
uvicorn api.main:app --reloadOpen http://127.0.0.1:8000
python3 -m http.server 5500Open http://127.0.0.1:5500/docs