Mining Association Rules in Hypertext Databases

In this work we propose a generalisation of the notion of association rule in the context of flat transactions to that of a composite association rule in the context of a structured directed graph, such as the world-wide-web. The techniques proposed aim at finding patterns in Che user behaviour when traversing such a hypertext system. We redefine the concepts of confidence and support for composite association rules, and two algorithms to mine such rules are proposed. Extensive experiments with random data were conducted and the results show that, in spite of the worst-case complexity analysis which indicates exponential behaviour, in practice the algorithms’ complexity, measured in the number of iterations performed, is linear in the number of nodes traversed.