Clustering XML Elements for Efficient Code Generation

Abstract

Several works [1,2,3] have been reported in literature to cluster XML documents and/or XML Schemas based on structure similarity, semantic similarity, edit distance similarity etc. Our current attempt clusters the elements presented in an XML Schema based on conversion cost that incurs to transform one object to the other. Our strategy would help a dynamic XML compiler to reproduce classes for new elements in C# programming language by modifying existing classes rather generating those from the very beginning.