Advantages of decision lists and implicit negatives in inductive logic programming

摘要: This paper demonstrates the capabilities offoidl, an inductive logic programming (ILP) system whose distinguishing characteristics are ability to produce first-order decision lists, use of output completeness assumption as a substitute for negative examples, and originally motivated by problem learning generate past tense English verbs; however, this its superior performance on two different sets benchmark ILP problems. Tests finite element mesh design show thatfoidl’s lists enable it generally more accurate results than range methods previously applied problem. with selection list-processing problems from Bratko’s introductory Prolog text demonstrate that combination implicit negatives intensionality allowfoidl learn correct programs far fewer examples thanfoil.