*A learner that makes no a priori assumptions regarding the identity of the target concept has no rational basis for classifying any unseen instances.*- So, what is the inductive bias?
- Given an algorithm $L$ and a set of training instances ${D}_{c}=\{\u27e8x,c(x)\u27e9\}$, let $L({x}_{i},{D}_{c})$ be the classification given to ${x}_{i}$ by $L$ after training on ${D}_{c}$. The inductive bias of $L$ is any minimal set of assertions $B$ for any target concept $c$ and examples $D$ s.t. $${\forall}_{{x}_{i}\in X}B\wedge {D}_{c}\wedge {x}_{i}\to L({x}_{i},{D}_{c})$$
- For example, the inductive bias of the Candidate-Elimination algorithm (with voting) is the assumption that the target concept is contained in the hypothesis space.
- That is, if we make that assumption then the
classification
*follows logically (by deduction)*.

30 of 31 |