Providing concise database covers instantly by recursive tile sampling

Sandy Moens; Mario Boley; Bart Goethals

doi:10.1007/978-3-319-11812-3_19

Providing concise database covers instantly by recursive tile sampling

Sandy Moens, Mario Boley, Bart Goethals

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Known pattern discovery algorithms for finding tilings (covers of 0/1-databases consisting of 1-rectangles) cannot be integrated in instant and interactive KD tools, because they do not satisfy at least one of two key requirements: a) to provide results within a short response time of only a few seconds and b) to return a concise set of patterns with only a few elements that nevertheless covers a large fraction of the input database. In this paper we present a novel randomized algorithm that works well under these requirements. It is based on the recursive application of a simple tile sample procedure that can be implemented efficiently using rejection sampling. While, as we analyse, the theoretical solution distribution can be weak in the worst case, the approach performs very well in practice and outperforms previous sampling as well as deterministic algorithms.

Original language	English
Title of host publication	Discovery Science - 17th International Conference, DS 2014, Proceedings
Editors	Sašo Džeroski, Panče Panov, Dragi Kocev, Ljupčo Todorovski
Publisher	Springer Verlag
Pages	216-227
Number of pages	12
ISBN (Electronic)	9783319118116
DOIs	https://doi.org/10.1007/978-3-319-11812-3_19
State	Published - 2014
Externally published	Yes
Event	17th International Conference on Discovery Science, DS 2014 - Bled, Slovenia Duration: 8 Oct 2014 → 10 Oct 2014

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	8777
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	17th International Conference on Discovery Science, DS 2014
Country/Territory	Slovenia
City	Bled
Period	8/10/14 → 10/10/14

Bibliographical note

Publisher Copyright:
© Springer International Publishing Switzerland 2014.

Keywords

Instant Pattern Mining
Sampling Closed Itemsets
Tiling Databases

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-319-11812-3_19

Cite this

Moens, S., Boley, M., & Goethals, B. (2014). Providing concise database covers instantly by recursive tile sampling. In S. Džeroski, P. Panov, D. Kocev, & L. Todorovski (Eds.), Discovery Science - 17th International Conference, DS 2014, Proceedings (pp. 216-227). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8777). Springer Verlag. https://doi.org/10.1007/978-3-319-11812-3_19

Moens, Sandy ; Boley, Mario ; Goethals, Bart. / Providing concise database covers instantly by recursive tile sampling. Discovery Science - 17th International Conference, DS 2014, Proceedings. editor / Sašo Džeroski ; Panče Panov ; Dragi Kocev ; Ljupčo Todorovski. Springer Verlag, 2014. pp. 216-227 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3fa1fa7a9b984301a85a559bdcff9be6,

title = "Providing concise database covers instantly by recursive tile sampling",

abstract = "Known pattern discovery algorithms for finding tilings (covers of 0/1-databases consisting of 1-rectangles) cannot be integrated in instant and interactive KD tools, because they do not satisfy at least one of two key requirements: a) to provide results within a short response time of only a few seconds and b) to return a concise set of patterns with only a few elements that nevertheless covers a large fraction of the input database. In this paper we present a novel randomized algorithm that works well under these requirements. It is based on the recursive application of a simple tile sample procedure that can be implemented efficiently using rejection sampling. While, as we analyse, the theoretical solution distribution can be weak in the worst case, the approach performs very well in practice and outperforms previous sampling as well as deterministic algorithms.",

keywords = "Instant Pattern Mining, Sampling Closed Itemsets, Tiling Databases",

author = "Sandy Moens and Mario Boley and Bart Goethals",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2014.; 17th International Conference on Discovery Science, DS 2014 ; Conference date: 08-10-2014 Through 10-10-2014",

year = "2014",

doi = "10.1007/978-3-319-11812-3_19",

language = "English",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "216--227",

editor = "Sa{\v s}o D{\v z}eroski and Pan{\v c}e Panov and Dragi Kocev and Ljup{\v c}o Todorovski",

booktitle = "Discovery Science - 17th International Conference, DS 2014, Proceedings",

address = "Germany",

}

Moens, S, Boley, M & Goethals, B 2014, Providing concise database covers instantly by recursive tile sampling. in S Džeroski, P Panov, D Kocev & L Todorovski (eds), Discovery Science - 17th International Conference, DS 2014, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8777, Springer Verlag, pp. 216-227, 17th International Conference on Discovery Science, DS 2014, Bled, Slovenia, 8/10/14. https://doi.org/10.1007/978-3-319-11812-3_19

Providing concise database covers instantly by recursive tile sampling. / Moens, Sandy; Boley, Mario; Goethals, Bart.
Discovery Science - 17th International Conference, DS 2014, Proceedings. ed. / Sašo Džeroski; Panče Panov; Dragi Kocev; Ljupčo Todorovski. Springer Verlag, 2014. p. 216-227 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8777).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Providing concise database covers instantly by recursive tile sampling

AU - Moens, Sandy

AU - Boley, Mario

AU - Goethals, Bart

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2014.

PY - 2014

Y1 - 2014

N2 - Known pattern discovery algorithms for finding tilings (covers of 0/1-databases consisting of 1-rectangles) cannot be integrated in instant and interactive KD tools, because they do not satisfy at least one of two key requirements: a) to provide results within a short response time of only a few seconds and b) to return a concise set of patterns with only a few elements that nevertheless covers a large fraction of the input database. In this paper we present a novel randomized algorithm that works well under these requirements. It is based on the recursive application of a simple tile sample procedure that can be implemented efficiently using rejection sampling. While, as we analyse, the theoretical solution distribution can be weak in the worst case, the approach performs very well in practice and outperforms previous sampling as well as deterministic algorithms.

AB - Known pattern discovery algorithms for finding tilings (covers of 0/1-databases consisting of 1-rectangles) cannot be integrated in instant and interactive KD tools, because they do not satisfy at least one of two key requirements: a) to provide results within a short response time of only a few seconds and b) to return a concise set of patterns with only a few elements that nevertheless covers a large fraction of the input database. In this paper we present a novel randomized algorithm that works well under these requirements. It is based on the recursive application of a simple tile sample procedure that can be implemented efficiently using rejection sampling. While, as we analyse, the theoretical solution distribution can be weak in the worst case, the approach performs very well in practice and outperforms previous sampling as well as deterministic algorithms.

KW - Instant Pattern Mining

KW - Sampling Closed Itemsets

KW - Tiling Databases

UR - http://www.scopus.com/inward/record.url?scp=84908689726&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-11812-3_19

DO - 10.1007/978-3-319-11812-3_19

M3 - Conference contribution

AN - SCOPUS:84908689726

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 216

EP - 227

BT - Discovery Science - 17th International Conference, DS 2014, Proceedings

A2 - Džeroski, Sašo

A2 - Panov, Panče

A2 - Kocev, Dragi

A2 - Todorovski, Ljupčo

PB - Springer Verlag

T2 - 17th International Conference on Discovery Science, DS 2014

Y2 - 8 October 2014 through 10 October 2014

ER -

Moens S, Boley M, Goethals B. Providing concise database covers instantly by recursive tile sampling. In Džeroski S, Panov P, Kocev D, Todorovski L, editors, Discovery Science - 17th International Conference, DS 2014, Proceedings. Springer Verlag. 2014. p. 216-227. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-11812-3_19

Providing concise database covers instantly by recursive tile sampling

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this