Bayesian Network Structure Learning Based on Rough Inclusion

Abstract

A kind of Bayesian network structure learning approach based on rough inclusion is put forward. First of all, the idea of the apriori algorithm is applied to mine frequent attribute sets through restraining support. Then, inclusion theory of rough set is used for mining cause and effect associated rules that determine arcs and their direction between Bayesian network variables. At one time, mining algorithm of associated rules and Bayesian network structure learning approach are presented. Finally, It shows rationality and validity of the approach by analyzing the applied procedure of example.