Improving the efficiency of ILP systems using an Incremental Language Level Search
Rui Carlos Camacho · 2002
We propose and evaluate a technique to improve the eciency of an ILP system. The technique avoids the generation of useless hypotheses. It denes a language bias coupled with a search strategy and is called Incremental Language Level Search (ILLS). The techniques have been encoded in the ILP system IndLog. The proposal leads to substantial eciency improvements in a set of ILP datasets referenced on the literature.