Logo des Repositoriums

Auflistung nach Autor:in "Pfeifle, Martin"

1 - 3 von 3
  • Konferenzbeitrag
    Efficient similarity search on vector sets
    (Datenbanksysteme in Business, Technologie und Web, 11. Fachtagung des GIFachbereichs “Datenbanken und Informationssysteme” (DBIS), 2005) Brecheisen, Stefan; Kriegel, Hans-Peter; Pfeifle, Martin
    Similarity search in database systems is becoming an increasingly important task in modern application domains such as multimedia, molecular biology, medical imaging, computer aided design and many others. Whereas most of the existing similarity models are based on feature vectors, there exist some models which use very complex object representations such as trees and graphs. A promising way between too simple and too complex object representations in similarity search are sets of feature vectors. In this paper, we first motivate the use of this modeling approach for complete object similarity search as well as for partial object similarity search. After introducing a distance measure between vector sets, suitable for many different ap- plication ranges, we present and discuss different filters which are indispensable for efficient query processing. In a broad experimental evaluation based on artificial and real-world test datasets, we show that our approach considerably outperforms both the sequential scan and metric index structures.
  • Konferenzbeitrag
    Measuring the quality of approximated clusterings
    (Datenbanksysteme in Business, Technologie und Web, 11. Fachtagung des GIFachbereichs “Datenbanken und Informationssysteme” (DBIS), 2005) Kriegel, Hans-Peter; Pfeifle, Martin
    Clustering has become an increasingly important task in modern application domains. In many areas, e.g. when clustering complex objects, in distributed clustering, or when clustering mobile objects, due to technical, security, or efficiency reasons it is not possible to compute an "optimal" clustering. Recently a lot of research has been done on efficiently computing approximated clusterings. Here, the crucial question is, how much quality has to be sacrificed for the achieved gain in efficiency. In this paper, we present suitable quality measures allowing us to compare approximated clusterings with reference clusterings. We first introduce a quality measure for clusters based on the symmetric set difference. Using this distance function between single clusters, we introduce a quality measure based on the minimum weight perfect matching of sets for comparing partitioning clusterings, as well as a quality measure based on the degree-2 edit distance for comparing hierarchical clusterings.
  • Konferenzbeitrag
    The paradigm of relational indexing: A survey
    (BTW 2003 – Datenbanksysteme für Business, Technologie und Web, Tagungsband der 10. BTW Konferenz, 2003) Kriegel, Hans-Peter; Pfeifle, Martin; Pötke, Marco; Seidl, Thomas
    In order to achieve efficient execution plans for queries comprising userdefined data types and predicates, the database system has to be provided with appropriate index structures, query processing methods, and optimization rules. Although available extensible indexing frameworks provide a gateway to seamlessly integrate user-defined access methods into the standard process of query optimization and execution, they do not facilitate the actual implementation of the access method itself. An internal enhancement of the database kernel is usually not an option for database developers. The embedding of a custom block-oriented index structure into concurrency control, recovery services and buffer management would cause extensive implementation efforts and maintenance cost, at the risk of weakening the reliability of the entire system. The server stability can be preserved by delegating index operations to an external process, but this approach induces severe performance bottlenecks due to context switches and inter-process communication. Therefore, we present in this paper the paradigm of relational access methods that perfectly fits to the common relational data model and is highly compatible with the extensible indexing frameworks of existing object-relational database systems.