Finding approximate patterns in undirected acyclic graphs

Document Type

Article

Publication Date

2-1-2002

Abstract

We consider an approximate pattern matching problem for undirected acyclic graphs. Specifically, let P be a pattern graph, D a data graph and t an integer. We present an algorithm to locate a subgraph in D whose distance from P is at most t. The distance measure used here is the degree-2 metric published previously. The time complexity of our algorithm is O(NpNDd √d log d) where Np and ND are the number of nodes in P and D, respectively; d = min dP,dD; dp and dD are the maximum degree of P and D, respectively. Central to our algorithm is a procedure for finding approximate patterns in rooted unordered trees and freely allowing cuts. We discuss two applications of the algorithms in chemical information search and website management on the Internet. © 2001 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.

Publication Title

Pattern Recognition

First Page Number

473

Last Page Number

483

DOI

10.1016/S0031-3203(01)00055-3

This document is currently not available here.

Share

COinS