Output details

How to find simple and accurate rules for viral protease cleavage specificities.

<28> Bioinformatics. The significance of the paper is to show the effectiveness of a previously proposed rule extraction method (Etchells and Lisboa ‘Orthogonal search-based rule extraction OSRE from trained neural networks: a practical and efficient approach’ IEEE Transactions on Neural Networks, 2006) in a rigorous validation with a highly challenging problem, Viral Protease Cleavage Specificity. The benchmark against the best alternative models for rule extraction demonstrated the originality of the method, which is to derive simple rules, in a disjunctive normal form with multivariate low-order terms, while maintaining classification accuracy.