Universal switching and side information portfolios under transaction costs using factor graphs

We consider the sequential portfolio investment problem. We demonstrate that the insights of Blum and Kalai's transaction costs algorithm may be used to construct more sophisticated algorithms. In particular, we show that transaction costs can be taken into account in Cover and Ordentlich's side information portfolio and Kozat and Singer's switching portfolio. For these, we present the corresponding universal (low regret) performance bounds for each of these portfolios. We then present factor graph representations of the algorithms and demonstrate that computationally efficient algorithms may be derived from the graphs. Finally, we present results of simulations of one of the derived algorithms and compare it to other portfolios.