An intrusion detection algorithm based on data streams mining and cognitive computing

With the emergence of large-volume and high-speed streaming data, traditional techniques for mining closed frequent itemsets has become inefficient. Online mining of closed frequent itemsets over streaming data is one of the most important issues in data streams minging. In view of the low efficiency of traditional closed frequent item data mining, a combined data structure based on the principle of cognitive computing is proposed, that is, combining the effective bit first with the extended dictionary frequent item list to form a mixed data structure that can identify the closed frequent information in data streams. At the same time, a variety of pruning strategies based on cognitive computing are proposed to avoid the generation of a large number of intermediate itemsets and to remove the non closed frequent term sets from the Hash Table of Closed Itemsets (CIHT). Closed Frequent Itemset Deletion and Search Strategy (CFIDWSS) is used to effectively add or remove the closed frequent itemsets, so as to greatly reduce the search space and improve the user response speed. The proposed algorithm solves the problem of low efficiency of data streams mining of closed frequent items. On the basis of the above algorithms, this paper proposes a new intrusion detection model. Through the mining of normal or abnormal patterns of data stream information, the corresponding database of network access pattern is established. Then the database is used to detect the intrusion online in real time and improve the detection accuracy of the system. Theoretical and experimental results show that the proposed algorithm and intrusion detection system have good performance.