Study On Sequence Patterns Mining And Its Application In Intrusion Detection

Sequence pattern mining is mining patterns based on time or other high-frequency ranking. Sequence pattern mining is an important research problem in data mining, for having a very broad range of applications, which includes customer-consume-behavior analysis, network- accessing-mode analysis, the analysis of scientific experiments, the early diagnosis of diseases, the forecast of natural disasters, the deciphering of the DNA sequence and so on. In this paper, the sequence pattern mining algorithms are deeply studied, and the achieved certain results of the research as follow:1. Sum up the development of sequence pattern mining algorithms from the beginning to now, analyze problems in sequence pattern mining, point out the advantage and disadvantage of various algorithms, and propose future development direction of sequence pattern mining algorithms.2. Propose a sequence pattern mining algorithm based on data stream. The method design a tiered storage structure to preserve the intermediate information as much as possible, thereby overcoming the flaws such as large flow of data, only one time visit. Experimental results show that the algorithm has good results, not only in mining sequence patterns effectively, but also increase the overall speed of mining sequence patterns to some extent.3. Propose a distributed sequence patterns mining algorithm based on the projection database. The algorithm based upon broad first, mining the overall sequence patterns with effectively saving disk space caused by projection database. The experimental results show that DLMSP algorithm is feasible and effective.4. Propose an intrusion detection algorithm based on the distributed sequence patterns mining, this algorithm is built on a distributed multi-agent system structures without center. We use a two-step detection algorithm which is based on a single agent detection cooperating with multi-agent detection to realize distributed detection.5. Implement an intrusion detection algorithm based on the distributed sequence patterns mining on a basis of predecessor's distributed intrusion detection system model based on agent using JAVA. Experimental test results show that the algorithm can detect network attacks in effect.