●Tools and Solutions for Semantic Reconciliation KLICT-project: ChaSe: Chain Services in ICT (FY-077)
1.2 Semantic reconciliation challenge One of the major problems in developing these kind of supply chain information services is that the collection process has to deal with heterogeneous data sources, fragmentation or incompleteness of information and possibly out-of-order and time-shifted data. In order to offer end-users of the different services coherent views on this information, data has to be cleansed, harmonized, reconciled and combined. The most intriguing part of this information integration process is definitely to reconcile information with different semantic frames of reference. In the last decades, the focus of many enterprise information systems has been on connecting information sources via networks, and on transforming information structures automatically for different purposes, but understanding the information is still a task reserved for humans who are aware of the context in which particular information was generated. This highest level of integration is called semantic reconciliation (or semantic integration). For automated systems, semantic reconciliation is still a hard nut to crack.
8.1 Definition The following definition of semantic reconciliation is given within the scope of the ChaSe project. Semantic reconciliation (is the process which) endeavors to understand the meaning of relevant information items in data from heterogeneous data sources, and to bring these information items in correspondence with an existing semantic data model. These information items include ?but are not restricted to- names, nouns, verbs, codes, classifications, measures, and references. Both, information items and data are supposed to be structured in one or another form. This form is called a schema. The preceding chapters in this document provide more information on the syntactical and structural requirements to these schemas and the associated syntactical and structural integration challenges, issues and solutions. Note that nearly always the underlying syntactical and structural integration are a necessary part of the overall reconciliation process. Semantic schema integration deals with either the merging of different conceptual schemas or a mapping of a conceptual schema on a conceptual reference or master schema.