We propose a Bayesian sequential learning framework for high-dimensional asset al-locations under model ambiguity and parameter uncertainty. The model is estimated via MCMC methods and allows for a wide range of data sources as inputs. Employing the proposed framework on a large set of...