Fast algorithms for comparison of similar unordered trees

Abstract

We present fast algorithms for computing the largest common subtree (LOST) and the optimal alignment when two similar unordered trees are given. We present an O(4Kn) time algorithm for the LOST problem for rooted trees, where n is the maximum size of two input trees and K is the minimum number of edit operations to obtain LOST. We extend this algorithm to unrooted trees and obtain an O(K4Kn) time algorithm. We also show that the alignment problem for rooted and unordered trees of bounded degree can be solved in linear time if K is bounded by a constant. © Springer-Verlag 2004.