Commodity style-integration is appealing because by forming a unique long-short portfolio with simultaneous exposure to mildly correlated factors, a larger risk premium can be captured over time than with any of the underlying standalone styles. A practical decision that a commodity style-integration investor faces at each rebalancing time is the relative weight of the predictive- or sorting-signal that underlies each standalone style. The authors of this paper develop a new Bayesian optimized integration (BOI) method that accounts for estimation risk in the style-weighting decision. Focusing on the problem of a commodity investor that seeks exposure to the carry, hedging pressure, momentum, skewness, and basis-momentum factors, they demonstrate that the BOI portfolio outperforms not only a battery of parametric style-integrations motivated by the portfolio optimization literature, but also the highly effective equal-weight integrated portfolio. The findings survive the consideration of transaction costs, alternative commodity scoring schemes, and long estimation windows.