An information integration framework for e-commerce

H. Benetti, D. Beneventano, S. Bergamaschi, F. Guerra, M. Vincini
2002 IEEE Intelligent Systems  
T he Web has transformed electronic information systems from single, isolated nodes into a worldwide network of information exchange and business transactions. In this context, companies have equipped themselves with high-capacity storage systems that contain data in several formats. The problems faced by these companies often emerge because the storage systems lack structural and application homogeneity in addition to a common ontology. The semantic differences generated by a lack of
more » ... ontology can lead to conflicts that range from simple name contradictions (when companies use different names to indicate the same data concept) to structural incompatibilities (when companies use different models to represent the same information types). One of the main challenges for e-commerce infrastructure designers is information sharing and retrieving data from different sources to obtain an integrated view that can overcome any contradictions or redundancies. Virtual catalogs can help overcome this challenge because they act as instruments to retrieve information dynamically from multiple catalogs and present unified product data to customers. Instead of having to interact with multiple heterogeneous catalogs, customers can instead interact with a virtual catalog in a straightforward, uniform manner. This article presents a virtual catalog project called Momis (mediator environment for multiple information sources). Momis is a mediator-based system for information extraction and integration that works with structured and semistructured data sources. 1,2 Momis includes a component called the SI-Designer for semiautomatically integrating the schemas of heterogeneous data sources, such as relational, object, XML, or semistructured sources. Starting from local source descriptions, the Global Schema Builder generates an integrated view of all data sources and expresses those views using XML. Momis lets you use the infrastructure with other open integration information sys-tems by simply interchanging XML data files. Momis creates XML global schema using different stages, first by creating a common thesaurus of intra and interschema relationships. Momis extracts the intraschema relationships by using inference techniques, then shares these relationships in the common thesaurus. After this initial phase, Momis enriches the common thesaurus with interschema relationships obtained using the lexical WordNet system (www.cogsci.princeton.edu/wn), which identifies the affinities between interschema concepts on the basis of their lexicon meaning. Momis also enriches the common thesaurus using the Artemis system, 3 which evaluates structural affinities among interschema concepts. System architecture Like other integration projects, Momis relies on a semantic approach based on the conceptual schemaor metadata-of the information sources, and on the I 3 architecture 4 (see Figure 1) . The system consists of the following functional elements that communicate using the Corba standard: a common data model, data wrappers, and a mediator. Momis defines the common data model, ODM I3 , according to the ODL I3 language, which describes source schemas for integration purposes. Momis treats ODM I3 and ODL I3 as subsets of the corresponding languages in the ODMG specificationaccording to the proposal for a standard mediator language developed by the I 3 -POB working group. 5 In addition, ODL I3 introduces new constructors to support the semantic integration process. Momis uses One of the main challenges for e-commerce infrastructure designers is to retrieve data from different sources and create a unified view that overcomes contradictions and redundancies. Virtual catalogs, such as the Momis project, can help synthesize data and present it in a unified manner to the customer.
doi:10.1109/5254.988444 fatcat:jpqavoul55gphmcs2qky54y2li