Output details

Learning probabilistic logic models from probabilistic examples

<24>

Probabilistic ILP systems learn probabilistic logic programs from examples and background knowledge. Clauses in a probabilistic logic program contain associated probability labels, which support logical and probabilistic inference. This paper extended the main approach to Probabilistic ILP by providing a framework for learning from examples which contain an associated probability label. Experiments show increased predictive accuracy in real-world problems. The paper represents the first attempt to extend the Probabilistic ILP framework to allow probabilistic examples, and was invited by the Program Committee to contribute to the special issue associated with the International Conference on Inductive Logic Programming.