Research On Approach To The Generation Of Frequent Itemset Based On Attention And Its Application In Network Management Of OAS

Office automation system is an important subsystem of intelligent building. At thesame time to enhance the level of building intelligent, the unauthorized use of thenetwork reduces the operating efficiency of the OAS. We can prevent unauthorized useby identifying the unauthorized use from the network. Association rule discoverymethod can efficiently obtain non trivial behaviors of network visiting implied containsat network access log in OAS, it’s an effective way to manage unauthorized use of thenetwork behavior. Identification of the network unauthorized use behavior in OAS isbeneficial to active control the network services on network abuse, overuse orunauthorized use to improve the efficiency of OAS.The network access log of OAS is massive, we would face two problems whenusing the association rule discovery methods to min the user behavior data:1) the timeof mining is too long;2) Excessive lawful access behavior affect the efficiency ofmining. In this paper, in order to improve the efficiency to get the behavior of the usernetwork access implicit in access log of OAS, we explored the association rulesdiscovery process.First, in order to reduce the time of getting the behavior of the user networkaccess implicit in access log of OAS by using the association rule discovery methods,the research draws on the early selection model of attention as information filter, andthen gives an approach to the generation of user focused frequent itemsets based on theearly selection model of attention as information filter and Apriori algorithm. Wedefined the user focus to characterize network management focus, achieved the formsexpression of attention, we also defined the precise and the recall for illustrate theeffectiveness of the methods. It verified the effectiveness and fast features of theproposed method through experiments.Secondly, in order to overcome the intersection of users focused with the wholeitemsets getting from the association rules discovery progress is empty or transactionsize is too small for a negative impact on the user network access behavior acquired, we discussed the extension method of users focus. Experimental results show that theapproach to the generation of generation of extended user focused expansion offrequent itemsets could solves the problem of original transset’s excessive filtration,able to maintain a certain scale transset to get the frequent itemsets more efficiently.Finally, in Anhui Key Laboratory of Intelligent Building, on the basis of using theIPTraf software intercepts all data packet of network access during a period of time asthe network access log, using the proposed method for network access log analysis, toachieve the management focus from the network management, and then we canoptimize the iptables firewall rules according to management focus for Anhui KeyLaboratory of Intelligent Building, this promoted the legitimate use of laboratorynetworks in OAS.