Dynamic portfolio optimization with inverse covariance clustering

Market conditions change continuously. However, in portfolio investment strategies, it is hard to account for this intrinsic non-stationarity. In paper, we propose address issue by using the Inverse Covariance Clustering (ICC) method identify inherent market states and then integrate such into a dynamic optimization process. Extensive experiments across three different markets, NASDAQ, FTSE HS300, over period of ten years, demonstrate advantages our proposed algorithm, termed Clustering-Portfolio Optimization (ICC-PO). The core ICC-PO methodology concerns identification clustering from analytics past data forecasting future state. It therefore agnostic specific choice. By applying same technique on ICC temporal cluster, instead whole train period, show that one can generate portfolios with substantially higher Sharpe Ratios, which are statistically more robust resilient great reductions maximum loss extreme situations. This shown be consistent periods, methods selection assets. • Each exists supreme regime persistently high log-likelihood. Optimizing only yields superior performance. process integrated. performance significant, markets methods. covariance sparsification improves both optimization.