An Improved Algorithm of Apriori

Based on Apriori algorithm,an improved algorithm of Apriori is presented in this paper.The main ideas of the algorithm are:1) Count the probability of each item(I₁,I₂,…,I_m) by scanning the DB for the first time;2) Item I_k and I_m appear in one record is P_km.min(P_k,P_m) ≤P_km ≤P_k*P_m,3) Count the support count of the frequence itemsets by scanning the DB another time;4) Output the association rule from the frequence itemsets.The detailed algorithm and its sample are described in this paper.Lastly we compare it with Apriori algorithm.The best quality of the new algorithm is that it reduce the times of scanning DB.