19 Box 1: Decision theory and control theory Bayesian decision theory introduces a loss function L(h,α(d)) for the cost of making a decision α(d) when the input is d and the true hypothesis is h. It proposes selecting the decision rule α∗ (.) that minimiz (1993)

Venue:L(h,α(d)) = 1 if α(d) �= h and L(h,α(d)) = 0 if α(d) = h. Then the