The Beginning

1. Initialize to most specific. For each positive training instance: if attribute constraint doesn't satisfy: Generalize it. (4, 1, 9)
6. experience type, target function, learning algorithm, representation (6, 7)
8. improve over task T with respect to performance measure P based on experience E (8)

2. more exp., criteria to select h / -> inconsistent samples (5, 2, 2)
3. subset of hypotheses from H consistend will all training exps D, defined by S & G boundary (2, 2)
4. Dc ^xi >> L(xi, Dc) (>> inductive inference) (8, 6, 1)
5. instances describable by attribute value pairs, discrete valued target function, disjunctive hypo may be required, possibly noisy training data (11, 2, 2)
7. minimal set of assertions B such that: for all xi € X: B ^Dc ^xi |= L(xi, Dc) (|= deductive inference) (9, 4)
9. knowledge dicovery in databases, 1989 (3)
10. Select exp. at random. Calculate with learned function on current weights. Calculate error. Update weights w1 + epsilon*feature1*error (3, 6, 6, 4)

Get Revising