Title of article :
A new taxonomy of sublinear right-to-left scanning keyword pattern matching algorithms
Author/Authors :
Loek Cleophas، نويسنده , , Bruce W. Watson، نويسنده , , Gerard Zwaan، نويسنده ,
Issue Information :
ماهنامه با شماره پیاپی سال 2010
Abstract :
A new taxonomy of sublinear (multiple) keyword pattern matching algorithms is presented. Based on an earlier taxonomy by the second and third authors, this new taxonomy includes not only suffix-based algorithms, but also factor- and factor-oracle-based algorithms. In particular, we show how suffix-based (Commentz-Walter like), factor- and factor-oracle-based sublinear keyword pattern matching algorithms can be seen as instantiations of a general sublinear algorithm skeleton. During processing, such algorithms shift or jump through the text in a forward or left-to-right direction, and read backward or right-to-left starting from positions in the text, i.e. they read suffixes of certain prefixes of the text. They use finite automata for efficient computation of string membership in a certain language. In addition, we show shift functions defined for the suffix-based algorithms to be reusable for factor- and factor-oracle-based algorithms. The taxonomy is based on deriving the algorithms from a common starting point by adding algorithm and problem details, to arrive at efficient or well-known algorithms. Such a presentation provides correctness arguments for the algorithms as well as clarity on how the algorithms are related to one another. In addition, it is helpful in the construction of a toolkit of the algorithms.
Keywords :
finite automata , Algorithm taxonomies , Pattern matching
Journal title :
Science of Computer Programming
Journal title :
Science of Computer Programming