Research On Weighted Association Rule And Routh Set For The Application Of Intrusion Detection

With the rapid development of network technology,people's lives have been greatly changed.At the same time,the complicated network environment makes the difficulty of the network security management increased,and network security issues can bemore attentioned in the Intemet.The application and development of the Internet bring convenience to people,but at the same time,network security issues trouble people.Frequent malicious attacks make the network paralysis of large area,and seriously influence the quality of people access to the network.The main functions of the IDS(Intrusion Detection System)is to protect the system,and analysis and predict the behavior of users.IDS is actually a system to detect illegal behavior,which can monitor various activities on the network.If the attack is detected,the IDS will send out a warning message to managers.Faced with the threat of network system at present,the traditional static defense has been unable to meet the demand.At this point,the relatively active.intrusion detection technology is paid more attention to.Based on this background,this paper put forward a mining method based on Weighted Association Rule and Rough Set,which will be applied to intrusion detection.The main steps of this algorithm are as follows.First of all,because the rough set is to use mathematical knowledge to describe the object of study,which can provide a precise and effective mathematical methods,this article usesRough Set Theory to the attribute reduction,and remove redundant useless properties.So the reducted data sets can not only accurately represent the decision system,but also can improve the computing speed.Then,based on the classical Association Rule,the weighted association rules method hasbeen proposed.Because the importance of each attribute is different,andIf the importance is the same,the result may be less accurate.In order to distinguish the degree,this article makes the attribute weighted to achieve.The weight of attribute can be calculate by the upper approximation and lower approximation ofRough Set Theory.And the greater the attribute is,the more important value is.We redefine the concept of weighted support and weighted confidence,which equals the weight multiply by he original support and confidence.Finally,in the classification phase,in order to avoid conflict rules,which is,a data matches multiple categories of rules and be classifiedby mistake.This article uses the support,confidence and all-confidence to sortrules.All-confidence has been provedcan be one of good standards,which can be used to distinguish the interest degree.So this article can be more persuadedwith all-confidence.It can improve the problem that new data can be classified by matchingthe sorted rules.In the end of this article,we do some experiments to verify this algorithm in the paper,and analysis the performance of the algorithm comprehensively.Compared with other algorithmsin the references which are under the background of intrusion detection.Experiments show that the accuracy of the proposed algorithm is superior to the algorithms in the references,which is efficient and feasible.