1 / 29

Ontology Mapping

Ontology Mapping. I3CON Workshop PerMIS August 24-26, 2004 Washington D.C., USA Marc Ehrig Institute AIFB, University of Karlsruhe. Agenda. Motivation Definitions Mapping Process Efficiency Evaluation Conclusion. Motivation. Semantic Web Many individual ontologies

imani-roman
Download Presentation

Ontology Mapping

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. Ontology Mapping I3CON Workshop PerMIS August 24-26, 2004 Washington D.C., USA Marc Ehrig Institute AIFB, University of Karlsruhe

  2. Agenda • Motivation • Definitions • Mapping Process • Efficiency • Evaluation • Conclusion Ontology Mapping

  3. Motivation • Semantic Web • Many individual ontologies • Distributed collaboration • Interoperability required • Automatic effective mapping necessary Ontology Mapping

  4. Mapping Definition • Given two ontologies O1 and O2, mapping one ontology onto another means that for each entity (concept C, relation R, or instance I) in ontology O1, we try to find a corresponding entity, which has the same intended meaning, in ontology O2. • map(e1i) = e2j • Complex mappings are not addressed: n:m, concept-relation,… Ontology Mapping

  5. Agenda • Motivation • Definitions • Mapping Process • Efficiency • Evaluation • Conclusion Ontology Mapping

  6. Iterations Input Process Entity Pair Selection Features Similarity Aggregation Interpretation Output Ontology Mapping

  7. Features Object Vehicle hasOwner hasSpeed Boat Car Speed Owner 250 km/h Marc Porsche KA-123 Ontology Mapping

  8. Similarity Measure • String similarity • Object Similarity • Set similarity Ontology Mapping

  9. Similarity Rules Ontology Mapping

  10. Iterations Input Process Entity Pair Selection Features Similarity Aggregation Interpretation Output Ontology Mapping

  11. Combination • How are the individual similarity measures combined? • Linearly • Weighted • Special Function Ontology Mapping

  12. Interpretation • From similarities to mappings • Threshold • map(e1j) = e2j ← sim(e1j ,e2j)>t Ontology Mapping

  13. Thing Vehicle simLabel = 0.0 simSuper = 1.0 1.0 simInstance = 0.9 hasSpecification Automobile simRelation = 0.9 Speed simCombination = 0.7 Object Marc’s Porsche fast 0.7 Vehicle hasOwner 0.9 Boat 0.9 Owner Car hasSpeed Speed Marc Porsche KA-123 250 km/h Example Ontology Mapping

  14. Agenda • Motivation • Definitions • Mapping Process • Efficiency • Evaluation • Conclusion Ontology Mapping

  15. Critical Operations • Complete comparison of all entity pairs • Expensive features e.g. fetching of all (inferred) instances of a concept • Costly heuristics e.g. Syntactic Similarity Ontology Mapping

  16. Assumptions • Complete comparison unnecessary. • Complex and costly methods can in essence be replaced by simpler methods. Ontology Mapping

  17. Reduction of Comparisons • Random Selection • Closest Label • Change Propagation • Combination Ontology Mapping

  18. Removal of Complex Features direct subclassOf all subclassOf direct instances all instances Ontology Mapping

  19. Complexity • c = (feat + sel + comp · (Σk simk + agg) + inter) · iter • NOM c = O((n + n2 + n2 ·(log2(n) + 1) + n) ·1) = O(n2 · log2(n)) • PROMPT c = O((n + n2 + n2 ·(1 + 0) + n) ·1) = O(n2) • QOM c = O((n + n·log(n) + n ·(1 + 1) + n) ·1) = O(n · log(n)) Ontology Mapping

  20. Agenda • Motivation • Definitions • Mapping Process • Efficiency • Evaluation • Conclusion Ontology Mapping

  21. Scenarios • Travel domain: Russia • 500 entities • Manual assigned mappings by test group Ontology Mapping

  22. 1,2 Label Sigmoid 1 0,8 n o i s i 0,6 c e r p 0,4 0,2 0 1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341 361 mapping with n highest similarity Precision Ontology Mapping

  23. 0,9 0,8 0,7 0,6 0,5 l l a c e r 0,4 Label 0,3 Sigmoid 0,2 0,1 0 1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305 324 343 362 mapping with n highest similarity Recall Ontology Mapping

  24. 0,9 0,8 0,7 0,6 e r 0,5 u s a e m 0,4 - f Label 0,3 Sigmoid 0,2 0,1 0 1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341 361 mapping with n highest similarity F-measure Ontology Mapping

  25. Efficiency Ontology Mapping

  26. Agenda • Motivation • Definitions • Mapping Process • Efficiency • Evaluation • Conclusion Ontology Mapping

  27. Conclusion • Automatic mappings are necessary. • Semantics help to determine better mappings. • Efficient approaches needed as ontology numbers and size increase. • Complexity of measures can be reduced. • Number of mapping candidates can be reduced. • Loss of quality is marginal. • Good increase in efficiency. Ontology Mapping

  28. Outlook • Machine learning to adapt to dynamically changing ontology environments • Increase evaluation basis • Addition of background knowledge e.g. WordNet • Integration into ontology applications e.g. for merging Ontology Mapping

  29. Thank you. Ontology Mapping

More Related