A Research of CM-Apriori Algorithm Based on Cluster Matrixes
Yangyang Li · Journal of Inner Mongolia University · 2008
To overcome the shortcomings of the Apriori algorithm,a new algorithm of CM-Apriori based on cluster matrix is proposed.This algorithm scans the transaction database only once and the cluster matrixes are created according to the item count.The frequent itemsets are generated by partial cluster matrixes.The CM-Apriori algorithm reduces the times of scanning transaction database and computation cost,thus improves the efficiency of association rule algorithm greatly.The example shows that this algorithm is an effective method of association rule mining.