Evolution of XPath Lists for Document Data Selection

摘要

XML has became a standard for structured data, and very often transformations from one specific format to another are needed. XSLT stylesheets are programs designed for this purpose, and they use XPath expressions to select sets of nodes within the document. In this paper a new version of an evolutionary algorithm that creates XSLT from examples is presented, improving on previously obtained results by testing a new individual representation with a new set of operators, based mainly on evolution of XPaths with a fixed XSLT program structure. The experiments show that this new representation, and a lower set of operators, yield better results in less generations that in our previous version.