Web Structure Mining using Learning Automata and Probabilistic Grammar

One of the data mining techniques on the web is web usage mining. The purpose of this mining, extracting useful information from the user interaction data that can be obtained when using the web. By extracting this information, we can determine the relationship between the web pages and then we can do operations such as clustering and ranking web pages. The purpose of this article is to determine the relationship between web pages, using probabilistic grammar and learning automata. In the proposed algorithm first probabilistic grammar generated and association rules are extracted from the user navigation paths stored in log files. Extracted association rules evaluated by learning automata and the relationship between web pages by learning automata is determined. The simulation results compared with Antweb, Bollen and E-DLA shows that the proposed method has a considerable efficiency.