WM+: an optimal multi-pattern string matching algorithm based on the WM algorithm

Authors:
Xunxun Chen;Binxing Fang;Lei Li;Yu Jiang
Affiliations:
Research Center of Computer Networks and Information Security Technology, Harbin Institute of Technology, Harbin, P.R.C.;Research Center of Computer Networks and Information Security Technology, Harbin Institute of Technology, Harbin, P.R.C.;Research Center of Computer Networks and Information Security Technology, Harbin Institute of Technology, Harbin, P.R.C.;Research Center of Computer Networks and Information Security Technology, Harbin Institute of Technology, Harbin, P.R.C.
Venue:
APPT'05 Proceedings of the 6th international conference on Advanced Parallel Processing Technologies
Year:
2005

Citing 4
Cited 1

Fast string matching using an n-gram algorithm

Software—Practice & Experience
A fast string searching algorithm

Communications of the ACM
Efficient string matching: an aid to bibliographic search

Communications of the ACM
Average-optimal multiple approximate string matching

CPM'03 Proceedings of the 14th annual conference on Combinatorial pattern matching

Exact online two-dimensional pattern matching using multiple pattern matching algorithms

Journal of Experimental Algorithmics (JEA)

Quantified Score

Hi-index	0.00

Visualization

Abstract

The WM algorithm, designed by Sun Wu and Udi Manber, is considered the fastest multi-pattern string matching algorithm in practice except when the pattern number is very large or the alphabet size is small[2]. Theoretically, the scanning time of WM is average-optimal (i.e. O(nlogσ(rm)/m)), but in the worst case, its scanning time can not be evaluated at all. The maximum shift of the original WM algorithm is m-B+1, where m is the minimum length of all patterns and B is the q-gram size. The tuned WM algorithm (abbreviated as WM+) can reach higher performance by improving the shift table building algorithm and combining the AC algorithm with the original WM algorithm. And the scanning time of the WM+ algorithm in the worst case is predictable. Experiments show that the scanning time of the WM+ algorithm is less or not great than that of the WM algorithm for varied size of m and number of patterns, especially in the worst case.