INDUCTIVE CONCEPT LEARNING The learning task

• Given:
• a language of examples , defining a set of a priori acceptable examples
• a language of concept descriptions , defining a set of a priory acceptable hypotheses
• a covers relation between  and , defining when an example e is covered by a hypothesis: covers(H, e), or covers(B, H, e)
• set of training examples P and N described in 
• quality criterion defined on strings from 
• Find:
• hypothesis H from , satisfying the quality criterion.

Standard quality criterion:

• completeness: H covers all positive examples


• consistency: H does not cover any neg. example