Using Label Information in a Genetic Programming Based Method for Acquiring Block Preserving Outerplanar Graph Patterns with Wildcards

Abstract

Machine learning and data mining from graph structured data have gained much attention. Many chemical compounds can be expressed by outerplanar graphs. We propose a method for acquiring characteristic block preserving outerplanar graph patterns with wildcards for vertex and edge labels, from positive and negative outerplanar graph data, by Genetic Programming using label connecting information of positive examples. We report experimental results on real chemical compound data and synthetic data.