An improved Apriori algorithm

Apriori algorithm has some abuses, such as too many scans of the database, large load of system´s I/O and vast unrelated middle itemsets. This paper proposes an improved Apriori algorithm to overcome these abuses. The improved algorithm reduces the set of candidates and accelerated the speed of the algorithm by adding the interest items. Breaking the traditional steps of the algorithm to reduce the database scans and bring down the load of system´s I/O. The algorithm improves the readability of the strong association by constructing the model of the interest measure. Experimental results show that the algorithm can improve the speed and efficiency of operation effectively.