MPO Gradient Descent

Project Description

This repository implements Multi-Objective Portfolio Optimization (MPO) using Gradient Descent (GD). It includes case studies for maximizing Sharpe ratio, minimizing Conditional Value at Risk (CVaR), and other portfolio optimization metrics. The project also explores functions like SparseMax and SoftMax and rounded versions of several activation functions.

Folder Structure

# Data folder containing input datasets
data/

# Results folder containing output images and metrics
results/

# Source code folder containing scripts and notebooks
src/

Installation

1. Clone the repository:

   git clone https://github.com/pventura1976/mpo-gradient-descent.git
   

2. Navigate to the project directory:

   cd mpo-gradient-descent
   

3. Install the required Python packages:

   pip install -r requirements.txt
   

Usage

• Notebooks: The src folder contains Jupyter notebooks for each case study. Open and run them using Jupyter Notebook or Jupyter Lab.

Authors

This project was developed by:

• Christian Oliva Moya
• Pedro Ramón Ventura Gómez
• Luis Lago Fernández

License

This project is licensed under the MIT License. See the LICENSE file for details.