A Comparison of ILP and Propositional

This paper presents an experimental comparison of two Inductive Logic Programming algorithms, Progol and Tilde , with C4.5, a propo-sitional learning algorithm, on a propositional dataset of road traac accidents. Rebalancing methods are described for handling the skewed distribution of positive and negative examples in this dataset, and the relative cost of errors of commission and omission in this domain. It is noted that before the use of these methods all algorithms perform worse than majority class. On rebalancing, all did signiicantly better. The conclusion drawn from the experimental results is that on such a propositional data set ILP algorithms perform competively in terms of predictive accuracy with propositional systems, but are signiicantly outperformed in terms of time taken for learning.