String Matching with Swaps in a Weighted Sequence

Abstract

A weighted sequence is a string where a set of characters may appear in certain positions with respectively known probabilities of occurrence. We concentrate on the problem of pattern matching with swaps (swapped m atnhing) in a weighted sequence, that is to locate all the positions in the sequence where there exists a swapped match of the pattern that has probability of appearance greater than a predefined constant. In this case, the number of swaps is not limited. We present a method for reducing the problem of swapped matching in a weighted sequence to the problem in a set of maximal factors of the sequence. We then give a almost optimal solution that takes O(nlogmlogσ) time. © Springer-Verlag 2004.