Learning k-Occurrence Regular Expressions from Positive and Negative Samples.

Deterministic regular expressions (DREs) are a core part of XML schema languages such as DTD/XSD and are used in different kinds of applications. Presently the most powerful model to learn DREs is k-occurrence regular expressions (k-OREs for short). However, there has been no algorithms can learn k-OREs from positive and negative samples. In this paper, we propose an efficient and effective algorithm to learn k-OREs from positive and negative samples. Our algorithm proceeds as follows: (1) learning deterministic k-OA from positive and negative samples based on genetic algorithm; (2) converting the k-OA into optimum deterministic k-OREs.