Affiliations: [a] Department of Computing and Software, McMaster University, Canada. [email protected] | [b] Department of Computing and Software, McMaster University, Canada; School of Engineering & Information Technology, Murdoch University, Perth, WA, Australia. [email protected]
Correspondence:
[*]
Address for correspondence: Department of Computing and Software, McMaster University, Canada
Note: [†] Research supported in part by an NSERC Discovery Grant.
Abstract: We study a central problem of string processing: the compact representation of a string by its frequently-occurring substrings. In this paper we propose an effective, easily-computed form of quasi-periodicity in strings, the frequency cover; that is, the longest of those repeating substrings u of w, |u| > 1, that occurs the maximum number of times in w. The advantage of this generalization is that it is not only applicable to all strings but also that it is the only generalized notion of cover yet proposed, which can be computed efficiently in linear time and space. We describe a simple data structure called the repeating substring frequency array (ℛ𝒮ℱ array) for the string w, which we show can be constructed in 𝒪(n) time and 𝒪(n) space, where |w| = n. We then use ℛ𝒮ℱ to compute all the frequency covers of w in linear time and space. Our research also allows us to give an alternate algorithm to compute all non-extendible repeating substrings in w, also in 𝒪(n) time and space.
