Abstract
The common association rules mining methods in multiple databases were inefficient due either to larger amount of candidate itemsets for communication overhead or higher times of database scan. Based on discussing the relation between the concept of pruned concept lattice and the representation of frequent itemsets, the Closed Frequent Itemsets of pruned concept lattice was defined. UMPCL, an approximately association rules mining method in horizontally partitioned databases based on multiple PCLs, was proposed. The main ideas of this method are using a frequent concept to represent some few of frequent itemsets to decrease the number of frequent itemsets and rules, and using a slightly lower support for pruning concept lattices before been merged to decrease the size of exchanged messages. The theoretic analysis and experiment show that such method is efficient. Key Word: Data Mining; Association rules; Pruned concept lattice; Distributed