Skip to content

nmchaves/stock-portfolio-selection

Repository files navigation

Stock Portfolio Selection via "Expert Pooling using Exponential Weighting (EPEx)"

Team members: Nico Chaves, Dominic Delgado

This was a course project for Stanford EE 378A (Statistical Signal Processing) in Spring 2016. For extensive details about our work, please see the report titled "final_report_chaves_delgado.pdf" in this repo.

In short, we developed and tested an ensemble algorithm for stock portfolio selection. The following figure shows the performance of our algorithm (shown in purple) in terms of portfolio value over time. Each algorithm began with an equal amount of money, and we only permitted trading on S&P 500 stocks (shorts were permitted). Note that EPEx attains the best performance over the tested interval. Further note that EPEx attained the highest Sharpe ratio of all of the tested algorithms.

Performance of Different Portfolio Selection Strategies

About

Experiments studying ensemble methods for stock portfolio selection

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published