Learning Logic Programs with Local Variables from Positive Examples
We present a polynomial time algorithm to learn a rich class of logic programs (called one-recursive programs) from positive examples alone. This class of programs uses the divide-and-conquer methodology and contains a wide range of programs such as append, reverse, merge, split, delete, insertion-sort, preorder and inorder traversal of binary trees, polynomial recognition, derivatives, sum of a list of numbers and allows local variables.
KeywordsLocal Variable Logic Program Polynomial Time Algorithm Positive Data Unit Clause
Unable to display preview. Download preview PDF.
- 1.H. Arimura, H. Ishizaka and T. Shinohara (1992), Polynomial time inference of a subclass of context-free transformations, Proc. of COLT’92, pp. 136–143.Google Scholar
- 3.M.R.K. Krishna Rao (1996), A class of Prolog programs inferable from positive data, Proc. of ALT’96, Lecture Notes in Computer Science 1160, pp. 272–84.Google Scholar
- 4.M.R.K. Krishna Rao and A. Sattar (1999), Learning logic programs with local variables from positive examples, Technical Report, Griffith University.Google Scholar
- 5.L. Sterling and E. Shapiro (1994), The Art of Prolog, MIT Press.Google Scholar