Clustering algorithms for portfolio optimization
Meysam Madani - CEO at Arnika
Thu, 22-Sep-2021 / 14:30 / Link:


Investment is not just picking the best asset, where rebalancing and constructing an optimized (weighted) set of assets is supposed to be a hard and critical issue to manage risk and return. To answer the following question: "Which diversification of assets lead to lower risk and better opportunities?" Harry Markowitz (1952) introduced modern portfolio theory (MPT) that is based on the assumption that "the portfolio return is the summation of the returns on each individual asset and the risk of an asset is the standard deviation of the asset returns". This naturally leads to an optimization problem to maximize returns by taking on a quantifiable amount of risk. Nowadays, the nature of financial markets is drastically changed in size as well as nature to become more complex and more fast-pacing, hence, asking for more sophisticated means and tools to be handled and analyzed. In this presentation, we try to review some applications of clustering methods in this area of research and we report on their effectiveness in portfolio theory as an intersection of the disciplines: finance, optimization and machine learning.


Meysam Madani is the CEO of Arnika, an AI Company with a focus on capital market and corporate finance. On the academic side, M. Madani has served as a Post-doc at Shari University of Technology and as an assistant professor at Shahid Beheshti University as an expert in Data science, while on the practical side he has been active as an AI/Data/Tech consultant for more than 12 years.