Machine Learning for visual risk analysis and hedge fund selection

01.07.2018

Empirische Kapitalmarktforschung, Risk & Optimization

The Hedge Fund Journal

Autoren

Huber, Dr. Claus

Abstract

The Self-Organising Map (SOM) is a Machine Learning tool to identify similarities in high-dimensional data. We show how it can be applied for manager selection to build robust portfolios. We describe a method that utilises some of the natural features of the SOM, e.g., the ability to process non-linearity in hedge fund returns and its visualisation capabilities that can be deployed for risk analysis, i.e., avoid managers with similar risk profiles and identify managers with unique risk profiles. Resulting portfolios exhibit significantly enhanced return/risk metrics and drawdown measures.

Weiterführende Informationen