Naive Bayesian classifier within ILP-R

When dealing with the classiication problems, current ILP systems often lag behind state-of-the-art attributional learners. Part of the blame can be ascribed to a much larger hypothesis space which, therefore, cannot be as thoroughly explored. However, sometimes it is due to the fact that ILP systems do not take into account the probabilistic aspects of hypotheses when classifying unseen examples. This paper proposes just that. We developed a naive Bayesian classiier within our ILP-R rst order learner. The learner itself uses a clever RELIEF based heuristic which is able to detect strong dependencies within the literal space when such dependencies exist. We conducted a series of experiments on artiicial and real-world data sets. The results show that the combination of ILP-R together with the naive Bayesian classiier sometimes signiicantly improves the classiication of unseen instances as measured by both classiication accuracy and average information score.