STUDY ON NETWORK INTRUSION DETECTION BASED ON IMPROVED APRIORI ALGORITHM

Association rule mining can effectively improve the performance of intrusion detection system, and generating frequent item-sets is a critical step in association rule mining. Through the research on Apriori algorithm, I propose an algorithm based on set and bit operation for mining frequent item-sets, and use digital view to express the transaction which used each item, use bit operating in digital view to calculate the number of the support of each item-set, and then avoid the problem of repeatedly scanning the database in Apriori algorithm. Experiments show that the algorithm has good performance in network intrusion detection.