Abstract:

In view of the disadvantages of the classic Apriori algorithm with multiple scan the database and generation of redundant candidate, we proposed an improved VMApriori algorithm. The algorithm combined with transaction data vector matrix and row candidate vectors representation method. We used
 quick sort of frequent item sets by individual frequency ascending rearrangement to enhance the efficiency of the algorithm. The experimental results show that the improved VMApriori algorithm can mining association rules correctly, and greatly improve the efficiency of execution.

Key words: VMApriori algorithm, association rule, itemsets optimization, vector matrix, data mining