Efficient Tree Mining Using Reverse Search

In this paper, we review our data mining algorithms for discovering frequent substructures in a large collection of semi-structured data, where both of the patterns and the data are modeled by labeled trees. These algorithms, namely FREQT for mining frequent ordered trees and UNOT for mining frequent unordered trees, efficiently enumerate all frequent tree patterns without duplicates using reverse search, which is a general scheme for designing efficient algorithms for hard enumeration problems, and incrementally compute of the occurrences of a pattern. We also discuss classes of trees to which reverse search is applicable, such as itemsets, sequential episodes, path trees, and graphs. Correspondence: Tatsuya Asai Department of Informatics, Kyushu University, 6-10-1 Hakozaki Higashi-ku, Fukuoka 812-8581, JAPAN E-mail: [email protected] phone: +81-92-642-2697, fax: +81-92-642-2698