Considering the characteristics of student achievement data is very huge, an improved idea of Apriori algorithm was proposed. First of all, change the storage of candidate itemsets and frequent itemsets, use linear tables to store itemsets, items of the minimum support, TID and the support of itemsets; then, scan transactions which items of the minimum support belong to, but not all, get the support of itemsets, thus reduce time consumption of producing the support of candidate itemsets; finally, find out frequent itemsets and association rules...