DeepFolio is a Python library for real-time portfolio optimization built on top of Google’s TensorFlow platform. Its design and development is based on the Portfolio Transformer (PT), a novel end-to-end portfolio asset allocation framework, inspired by the numerous successes of attention mechanisms in natural language processing. With PT’s full encoder-decoder architecture,specialized time encoding layers, and gating components, it has a high capacity to learn long-term dependencies among portfolio assets and hence can adapt more quickly to changing market conditions. Deepfolio combines optimization techniques (both convex and non-convex) with deep learning approaches to provide a powerful toolkit for investment professionals and researchers.

Features

• Portfolio Transformer: Attention-Based Asset Allocation Framework

• Differentiable portfolio optimization
• Real-time optimization
• Robust and multi-period optimization
• Multi-asset class support
• Backtesting system
• Risk management tools
• Factor model integration
• Automated hyperparameter tuning (Backed by Optuna)
• Trade execution simulation
• Event-driven rebalancing
• Comprehensive reporting
• Sentiment analysis integration
• Tax-aware optimization
• Interactive visualization dashboard

Installation

pip install --upgrade deepfolio

Quick Start

from deepfolio.models import DiffOptPortfolio
from deepfolio.optimizers import CustomOptimizer
from deepfolio import Backtester

# Initialize the model
model = DiffOptPortfolio(input_dim=50, n_assets=10, hidden_dim=64)

# Create an optimizer
optimizer = CustomOptimizer(model.parameters())

# Load your data
features, returns = load_your_data()

# Create a backtester
backtester = Backtester(model, {'features': features, 'returns': returns})

# Run backtesting
backtester.run()

# Get results
results = backtester.get_results()
print(f"Sharpe Ratio: {results['sharpe_ratio']}")
print(f"Max Drawdown: {results['max_drawdown']}")

Advanced Usage

Portfolio Transformer : Attention-Based Asset Allocation

from deepfolio.optimizers import portfolio_transformer
input_shape = (30, 10)  # 30 time steps, 10 assets
d_model = 64
num_heads = 4
dff = 128
num_layers = 4

model = portfolio_transformer(input_shape, d_model, num_heads, dff, num_layers)
model.compile(optimizer='adam', loss='mse')  # Adjust loss as per the paper's requirements
model.summary()

Real-time Optimization

from deepfolio.models import RealtimeOptimizer
from deepfolio.data import DataSource

data_source = DataSource(api_key="your_api_key")
optimizer = RealtimeOptimizer(model, data_source)
optimizer.start()

Multi-Asset Optimization

from deepfolio.models import MultiAssetDiffOptPortfolio

asset_classes = ['stocks', 'bonds', 'commodities']
input_dims = {'stocks': 50, 'bonds': 30, 'commodities': 20}
hidden_dims = {'stocks': 64, 'bonds': 32, 'commodities': 32}
model = MultiAssetDiffOptPortfolio(asset_classes, input_dims, hidden_dims)

Tax-Aware Optimization

from deepfolio.optimizers import TaxOptimizer

tax_optimizer = TaxOptimizer()
optimal_trades = tax_optimizer.optimize(current_portfolio, target_weights, prices, cost_basis, holding_period)

Interactive Dashboard

from deepfolio.utils import PortfolioDashboard

dashboard = PortfolioDashboard(portfolio_data, benchmark_data)
dashboard.run()

Documentation

For detailed documentation, please visit our documentation site.

Contributing

We welcome contributions! Please see our contributing guidelines for more details.

License

This project is licensed under the BSD-2-Clause License- see the LICENSE file for details.

Reference

[1] Damian Kisiel, Denise Gorse (2022). Portfolio Transformer for Attention-Based Asset Allocation arXiv:2206.03246 [q-fin.PM]

Acknowledgments

• This package leverages the power of TensorFlow for efficient portfolio optimization.
• Thanks to the financial machine learning community for inspiring many of the implemented methods.