1 / 10

WG 3: Data Integration

WG 3: Data Integration. Louiqa, Alejandro, Vincent, Monica, Stefan, Frank, Gerd, Felix. Data Integration. Architectures and application domains Quality criteria and issues specific to integration Compared 2 data sources based on quality criteria.

tosca
Télécharger la présentation

WG 3: Data Integration

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. WG 3: Data Integration Louiqa, Alejandro, Vincent, Monica, Stefan, Frank, Gerd, Felix

  2. Data Integration • Architectures and application domains • Quality criteria and issues specific to integration • Compared 2 data sources based on quality criteria. • Characterize data integration (2 sources) based on quality criteria.

  3. Architectures and Domains • Architectures • Materialized integration (data warehouse) • Virtual integration (mediator/wrapper) • Materialized solution provides more opportunities for improving quality OFFLINE. • Application domains • Bibliographic

  4. Quality Criteria • Duplicates • Inherited (already in sources) • Through integration • Synonyms and homonyms • Granularity • Higher granularity typically means higher level of detail and expressive power of queries. • Potentially lowest common denominator for result. • Completeness • Object cardinality • Attribute cardinality • CWA not applicable (open to discussion) • Content description of and knowledge about a source

  5. Quality Criteria 2 • Ontologies • Conceptual schema is a sub-concept of ontology • Databases described by schemata • Integration possible without ontologies, • but easier and better with them. • Quality of ontologies versus quality of data • High quality of schema => High quality of data • High quality of schema => Ease of integration

  6. Quality Criteria 3 • Currency • Timeliness, freshness, up-to-dateness,… • Materialization => Reduced currency • Virtual integration => Increased currency • Data warehouse: out-of-date data can be interpreted as missing data. • Response Time (Time, Cost, Delay) (later) • Greater impact on virtual integration • Trade-offs and user interests • Availability • DW increases availability

  7. Compare 2 sources based on DQ • Citeseer • Domain: All online CS papers • and DBLP • Domain: All DB- and LP- and algorithmic papers • Duplicates • DBLP eliminated more duplicates • DBLP had higher granularity, e.g., pubtype • Citeseer is automated => more duplicates • Completeness • Citeseer higher object and attribute cardinality • Citeseer: 600,000 publications • DBLP: 420,000 publications • Ontologies • Both have schemata • Currency … • DBLP: Manual updates

  8. Improved quality through integration? • Duplicates + Potential to improve (reduce) duplicates using combined data sources. - Potential of introducing new duplicates • Granularity - Structured model may force choice of lowest common denominator (month vs. day). Depends on integrating operator and model. • Completeness + Typically improves • Degree depends on level of object / attribute overlap

  9. Improved quality through integration? • Ontologies - Comparing/merging ontologies typically requires a trade-off that reduces quality. + With increased effort, quality of combined ontology (integration) may improve. • Currency • An old value is not necessarily an out-of-date value. • Application specific • Stock quotes with different up-to-date values: retain all values? • Address information: Delete out-of-date information. + Integrated result has a guaranteed currency (lower bound).

  10. Improved quality through integration? • Response Time (Time, Cost, Delay) - Greater impact on virtual integration • Trade-offs and user interests • Availability + DW increases availability • Depends on degree of overlap • Depends on specific query • Type of operator: Union (partial) or Join (fail)

More Related