Graph transformation to infer schemata from XML documents

Semi-structured data are characterized by the lack of a predefined schema. This heterogeneity simplifies the management of such data, but analysis and queries become more difficult and demand for schemata that describe these data. Super-imposed structures cannot be as general as predefined ones, but ease the retrieval of the information embedded in such data. The paper adopts XML as the language to render semi-structured data and proposes an approach - based on graph transformation techniques - to infer the schemata of XML documents.