The generic rough set inductive logic programming model and motifs in strings

Abstract

The gRS–ILP model (generic Rough Set Inductive Logic Programming model) provides a framework for Inductive Logic Programming when the setting is imprecise and any induced logic program will not be able to distinguish between certain positive and negative examples. However, in this rough setting, where it is inherently not possible to describe the entire data with 100% accuracy, it is possible to definitively describe part of the data with 100% accuracy. The gRS–ILP model is extended in this paper to motifs in strings. An illustrative experiment is presented using the ILP system Progol on transmembrane domains in amino acid sequences.