Period recovery over the hamming and edit distances

Abstract

A string S of length n has period P of length p if S[i] = S[i+p] for all 1 ≤ i ≤ n−p and n ≥ 2p. The shortest such substring, P, is called the period of S, and the string S is called periodic in P. In this paper we investigate the period recovery problem. Given a string S of length n, find the primitive period(s) P such that the distance between S and the string that is periodic in P is below a threshold τ. We consider the period recovery problem over both the Hamming distance and the edit distance. For the Hamming distance case, we present an O(n log n) time algorithm, where τ is given as (Formula Presented), for 0 < ε < 1. For the edit distance case, (Formula Presented), and we provide an O(n4/3) time algorithm.