Data Mining and Efficient Similarity Retrieval of Hierarchical Structures from Large Databases
Hierarchical structures are to be found everywhere in biomedical applications, multimedia databases and spatio-temporal database applications. Tree edit distances (TED) have been proposed as a well-suited similarity model for hierarchical structures. Since they are highly computational expensive, efficiency has to be improved when applying TED to large databases of complex objects. We have developed several histogram-based filter distances that meet high filter qualities as they are complete, selective and efficient when employed in multi-step query processing architectures. Experiments demonstrate the good performance of our techniques.
|Authors:||Seidl T., Schönauer S., Kailing K.|
|Published in:||J.-R. Sack, M. Sester, M. Worboys, P. van Oosterom (eds.): Spatial Data - mining, processing and communicating. Dagstuhl Seminar 06101|
|Publisher:||Internationales Begegnungs- und Forschungszentrum für Informatik (IBFI), Schloss Dagstuhl, Germany - Dagstuhl, Germany|
|Forschungsgebiet:||Data Analysis and Knowledge Extraction|