Efficient Computing of Longest Previous Reverse Factors

We study the problem of finding the longest previous reverse factor occurring at each position of a string. This is a generalisation of a notion used for the optimal detection of various types of palindromes in a string. We describe two algorithms for computing the LPrF table of a string, one from its Suffix Tree and the second from its Suffix Automaton respectively. These algorithms run in linear time on a fixed size alphabet. Typical applications of these algorithms are for RNA secondary structure prediction and for text compression.