Design and Implementation of an Improved Apriori Data Mining Algorithm

摘要

The need to scan the database D many times when the Apriori algorithm is applied to a large database causes the I/O load overhead of the disk to increase. An improved Apriori algorithm is designed. After scanning the original database D for the first time, it generates a candidate transaction database A_k. In the process of generating frequent itemsets, the candidate transaction database A_k is scanned each time. Experiments show that if the K value is very large, the number of A_k will be much less than that of the original database, which will solve the problem of I/O overload and reduce operation time, so as to achieve the purpose of optimizing Apriori algorithm.