Large-Scale Mean-Variance Optimization and Chunking Algorithm

Abstract

We propose a new, highly effective and easy-to-implement algorithm for solving large-scale mean-variance optimization problems --- with weight upper bound constraints and without short sales --- when the size of mean-variance portfolios is much smaller than the number of assets, which is almost always the case. Our novel algorithm is built on the novel representation of mean-variance models in terms of the support vector data description --- an unsupervised machine learning algorithm designed for a one-class classification problem --- and the chunking algorithm, a decomposition algorithm for support vector machine.

Keywords: Mean-Variance Optimization, One-Class Classification, Machine Learning, Support Vector Machine, Support Vector Data Description, Chunking Algorithm, Quadratic Programming

JEL Classification: G11, D83, C61, C63