Intrusion Detection Method Based On Frequent Pattern

People rely on the network perpetually as the extensive application of it. TheE-commerce and online payment, which related to the money transaction appear in quantity.So the security of the network becomes more and more important. The intrusion detectionsystem works as the second firewall of the computers. It plays a significant role in computersecurity guarantees. The IDS is stepping forward with the advance of the network.This paper introduces a new algorithm of the application of data mining in IDS, and themining of the frequent pattern is an important branch of data mining. This paper presents thealgorithm of mining the frequent pattern and optimizes it.Frequent pattern is of great importance in data mining technology, but also it is thebottleneck of it. The research of the frequent pattern can improve the efficiency of datamining. This paper draws inspiration from Zengyou He’s isolated point detecting algorithmbased on frequent pattern. This paper raises the intrusion detection algorithm based on longfrequent pattern.However the research of the long frequent pattern mining is not adequate yet. This paperbrings forward a new measurement called normal degree. It mines the long frequent pattern inevery transaction, and sorts them in accordance with the number of the frequent pattern. Thenormal degree of every transaction can be received. The transaction with ND less than theleast ND is likely to be the intrusion. This paper introduces the concrete algorithm, pseudocode and the algorithm architecture. Lastly, this paper proves that the efficiency of thedetection is largely improved with the experiments.