Variants of Validity and their Impact on the Overall Test Space

May 1998

Proceedings of the Eleventh International Florida Artificial Intelligence Research Society Conference (FLAIRS 1998)

Volume

Proceedings of the Eleventh International Florida Artificial Intelligence Research Society Conference (FLAIRS 1998)

All Papers

Dealing with the question whether or not a given system does suffice some interesting property one is confronted with the problem, how to navigate appropriately through the available knowledge space to prove or to refute that property under investigation. Moreover it is desireable to get an algorithm which solves that task efficiently. Applied to the area of system validation, we will propose some solution allowing for the reduction of test cases, i.e. of the scale of the knowledge space to be investigated, when vali-dating some target system by testing. The key idea for test case reduction is to exploit certain inheritance properties of the underlying space of input data. In its right perspective, inheritance is induction. Due to the impossibility to create any general induction scheme for deductive justification of inductive reasoning, there arises the necessity of domain-dependent variants.