An optimization problem related to bloom filters with bit patterns

Damaschke, Peter,1963Chalmers tekniska högskola,Chalmers University of Technology (författare)

© 2018, Springer International Publishing AG. Bloom filters are hash-based data structures for membership queries without false negatives widely used across many application domains. They also have become a central data structure in bioinformatics. In genomics applications and DNA sequencing the number of items and number of queries are frequently measured in the hundreds of billions. Consequently, issues of cache behavior and hash function overhead become a pressing issue. Blocked Bloom filters with bit patterns offer a variant that can better cope with cache misses and reduce the amount of hashing. In this work we state an optimization problem concerning the minimum false positive rate for given numbers of memory bits, stored elements, and patterns. The aim is to initiate the study of pattern designs best suited for the use in Bloom filters. We provide partial results about the structure of optimal solutions and a link to two-stage group testing.

Schliep, Alexander,1967Gothenburg University,Göteborgs universitet,Institutionen för data- och informationsteknik, datavetenskap (GU),Department of Computer Science and Engineering, Computing Science (GU),University of Gothenburg(Swepub:cth)schliep (författare)
Chalmers tekniska högskolaInstitutionen för data- och informationsteknik, datavetenskap (GU) (creator_code:org_t)

Ingår i:Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)Cham : Springer International Publishing0302-97431611-3349

Om ämnet

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.