FinRL Documentation

Welcome to the comprehensive documentation for FinRL - a Deep Reinforcement Learning library for financial markets!

🚀 What is FinRL?

FinRL is a powerful Python library that provides implementations of state-of-the-art deep reinforcement learning algorithms for financial applications, including:

Stock Trading: Multi-asset portfolio management
Cryptocurrency Trading: Real-time crypto trading strategies
Portfolio Optimization: Dynamic portfolio allocation
Risk Management: Advanced risk metrics and controls

🏗️ Architecture Overview

FinRL follows a three-layer architecture:

graph TD
    A[Data Layer] --> B[Environment Layer]
    B --> C[Agent Layer]
    C --> D[Trading Results]

Data Layer: Data collection, preprocessing, and feature engineering
Environment Layer: Trading environments (Gym-compatible)
Agent Layer: RL algorithms (A2C, PPO, DDPG, SAC, TD3)

⚡ Quick Start

from finrl.meta.preprocessor.yahoodownloader import YahooDownloader
from finrl.meta.preprocessor.preprocessors import FeatureEngineer
from finrl.meta.env_stock_trading.env_stocktrading import StockTradingEnv
from finrl.agents.stablebaselines3.models import DRLAgent

# 1. Download data
df = YahooDownloader(
    start_date="2020-01-01",
    end_date="2023-01-01", 
    ticker_list=["AAPL", "GOOGL", "MSFT"]
).fetch_data()

# 2. Process features
fe = FeatureEngineer(use_technical_indicator=True)
processed = fe.preprocess_data(df)

# 3. Create environment
env = StockTradingEnv(df=processed, **env_kwargs)

# 4. Train agent
agent = DRLAgent(env=env)
model = agent.get_model("ppo")
trained_model = agent.train_model(model, total_timesteps=50000)

📚 Documentation Sections

API Reference

Environment Classes - Trading environment APIs
Agent Classes - RL agent implementations

Data & Features

Data Format - Data requirements and format specifications

Advanced Features

Ensemble Methods - Multi-model strategies

Examples & Tutorials

Cryptocurrency Trading - Crypto trading strategies and examples

Troubleshooting

Common Issues - Solutions to frequent problems

🎯 Key Features

Multi-Algorithm Support

FinRL supports 5 state-of-the-art RL algorithms: A2C, PPO, DDPG, SAC, and TD3

Flexible Environments

Pre-built environments for stocks, crypto, and portfolio optimization

Performance Tracking

Comprehensive backtesting with 20+ financial metrics

Risk Management

Built-in turbulence detection and risk controls

🤝 Contributing

We welcome contributions! Please check our GitHub repository for contribution guidelines.

📄 License

This project is licensed under the MIT License.