Research on Association Rules Algorithm Based on Bit Storage and Deep Pruning

Association rule mining is an important part of data mining and is widely used in many fields. Mining frequent itemsets is the most important step and technology. In this paper, the Eclat algorithm in the frequent itemset mining algorithm is taken as the research point. In order to solve the problem that the Eclat algorithm increases the size of the itemset, causing the vertical list to store transaction records and intersection operations, etc, consumes a lot of time and memory. From the algorithm itself and storage mechanism considerations, a DBEclat (Deep Bit Eclat) algorithm based on bit storage mechanism and deep pruning strategy is proposed. The algorithm stores the vertical data list in a binary form that reduces memory consumption, reduces the time consumption of the intersection operation by using the AND operation of the binary data, and combines the multi-angle deep pruning strategy to compress the size of the candidate set. The experimental results show that the proposed algorithm has higher time efficiency and lower memory consumption than the original algorithm.