A filter method for the weighted local similarity search problem

Abstract

In contrast to the extensively studied k-differences problem, in the weighted local similarity search problem one searches for approximate matches of subwords of a pattern and subwords of a text whose lengths exceed a certain threshold. Moreover, arbitrary gap and substitution weights are allowed. In this paper, two new prefilter algorithms for the weighted local similarity search problem are presented. These overcome the disadvantages of a similar filter algorithm devised by Myers.