Processing Queries over Conceptual Integration Schemas

点击翻译按钮获取中文摘要

Business intelligence (BI) is a software industry term that denotes the use of information within organizations to allow deciders, usually executives, to make informed decisions, and to effectively run operations using known data. As a Computer Science research field, it encompasses several subfields of Computer Science, among which data integration (aka data warehousing) plays a pivotal role. Ideally, integration would happen almost automatically. One important aim of BI is to be end-user oriented: decision makers and business users need only to focus on modelling their data needs in (usually high-level ) business terms that they fully understand. A business user would specify what information is needed and in what form, and the integration system would satisfy the request by producing integrated data as automatically as possible. To achieve this, we must raise the level of abstraction significantly to accommodate business users. The latter would use conceptual-level schemas to model their integration needs. Once those schemas are in place, business users can pose queries expressed in a conceptual query language against those conceptual schemas. Building on previous work on conceptual integration schema modelling, the objectives of the present proposal are as follows: (1) address the problem of answering business queries over conceptual integration schemas; (2) identify and study tractable classes of queries that are of practical importance; (3) develop and implement efficient algorithms for some of those tractable classes.