DeepDow
A Python framework for portfolio optimization using deep learning to allocate investment weights in a single forward pass.
What is DeepDow?
deepdow is a Python framework that connects portfolio optimization with deep learning, enabling the creation of neural networks that perform investment weight allocation in a single forward pass. It merges forecasting and optimization steps into a fully differentiable pipeline, allowing researchers to train models end-to-end using gradient descent. The framework focuses on buy-and-hold strategies, simplifying the design of allocation networks without active trading complexities.
Researchers and data scientists in quantitative finance who want to experiment with deep learning for portfolio optimization, and developers building differentiable allocation models using PyTorch.
deepdow offers a unified, differentiable framework that replaces traditional multi-step portfolio optimization with a single neural network pipeline, providing flexibility, extensibility, and integration with tools like cvxpylayers for convex optimization. Its modular design and support for various loss functions make it ideal for rapid experimentation and research.
Overview
Portfolio optimization with deep learning.
Use Cases
Best For
- Researching deep learning approaches to portfolio weight allocation
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
