Abstract:
|
The tree matching problem is considered of given labeled
trees P and T, determining if the pattern tree P can be obtained from
the text tree T by deleting degree-one and degree-two nodes and, in
the case of unordered trees, by also permuting siblings. The constrained
tree inclusion problem is more sensitive to the structure of the pattern
tree than the general tree inclusion problem. Further, it can be solved in
polynomial time for both unordered and ordered trees. Algorithms based
on the subtree homeomorphism algorithm of (Chung, 1987) are presented
that solve the constrained tree inclusion problem in O(m1.5n) time on
unordered trees with m and n nodes, and in O(mn) time on ordered
trees, using O(mn) additional space. These algorithms can be improved
using results of (Shamir and Tsur, 1999) to run in O((m1.5/ logm)n)
and O((m/logm)n) time, respectively. |