Methods for mining frequent sequential patterns

Authors:
Linhui Jiang;Howard J. Hamilton
Affiliations:
Department of Computer Science, University of Regina, Regina, SK, Canada;Department of Computer Science, University of Regina, Regina, SK, Canada
Venue:
AI'03 Proceedings of the 16th Canadian society for computational studies of intelligence conference on Advances in artificial intelligence
Year:
2003

Citing 3
Cited 3

Principles of data mining

Principles of data mining
Mining Sequential Patterns

ICDE '95 Proceedings of the Eleventh International Conference on Data Engineering
A quick look at methods for mining long subsequences

AI'03 Proceedings of the 16th Canadian society for computational studies of intelligence conference on Advances in artificial intelligence

Unsupervised pattern mining from symbolic temporal data

ACM SIGKDD Explorations Newsletter - Special issue on data mining for health informatics
A quick look at methods for mining long subsequences

AI'03 Proceedings of the 16th Canadian society for computational studies of intelligence conference on Advances in artificial intelligence
Mining transposed motifs in music

Journal of Intelligent Information Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

We consider the problem of finding frequent subsequences in sequential data. We examine three algorithms using a trie with K levels. The O(K2n) breadth-first (BF) algorithm inserts a pattern into the trie at level k only if level k-1 has been completed. The O(Kn) depth-first (DF) algorithm inserts a pattern and all its prefixes into the trie before examining another pattern. A threshold is used to store only frequent subsequences. Since DF cannot apply the threshold until the trie is complete, it makes poor use of memory. The heuristic depth-first (HDF) algorithm, a variant of DF, uses the threshold in the same manner as BF. HDF gains efficiency but loses a predictable amount of accuracy.