Mining frequent close itemsets over data stream by transaction list group

Mining frequent itemsets is a basic task of the data stream mining.Recently many approximate algorithms can mine frequent itemsets over data stream.However,these algorithms still cannot efficiently reduce space and time cost.To improve the efficiency,mining frequent close itemsets over data stream is proposed to reduce the number of frequent itemsets.Referring to the algorithms of Relim and Manku,the transaction list group is imported as the synopsis data structure,and a new algorithm of mining frequent close itemsets is put forward.At the end,experiments are done to prove the efficiency of this algorithm.