New Criterion for Mining Strong Association Rules in Unbalanced Events

Association rules mining is an important task in data mining and the normal measures support and confidence are useful for finding association rules between the items. However, the process of finding frequent items would prune infrequent items which may include some useful relationships of association patterns. The new measures comsup, comcof and comsup´ are proposed to resolve this problem effectively. By comparison and taking examples, these new measures proved to be effective in the special situation, and some interesting rules could be found in the unbalanced events in which include the infrequent items.