Pattern Matching on Sparse Suffix Trees

Abstract

We consider a compact text index based on evenly spaced sparse suffix trees of a text [9]. Such a tree is defined by partitioning the text into blocks of equal size and constructing the suffix tree only for those suffixes that start at block boundaries. We propose a new pattern matching algorithm on this structure. The algorithm is based on a notion of suffix links different from that of [9] and on the packing of several letters into one computer word.