Research Of Database Intrusion Detection System Based On Improved Association Rules Algorithm

The rapid development of computer science makes data communion become more and more abroad. Meanwhile, the management of database system is more and more difficult. With intruders' developing skills and more complex attack tools, the only statistic defending mode is not sufficient. Intrusion Detection which can defend forwardly is the right technology to meet the needing. By detecting network and system interior data and activity, this technology can find the impossible intrusion activity and alarm or cut down the connection with the intruder. Intrusion Detection not only can keep away the exterior intrusion, but also can detect the interior illegal user. Data Mining is the common technology in Intrusion Detection these years. It can find the potential useful information to forecast the future so that it improved Intrusion Detection efficiency distinctly. So do research of Data Mining applying in managing Database Safety is theoretically and practically significant.This paper firstly summarized the database safety technology and present research actuality to indicate the safety threatens to database, and discussed the Intrusion Detection technology's important value in maintaining database safety, and comprised with several familiar Intrusion Detection technologies. Then related with Data Mining applying in Intrusion Detection, this paper indicated the advantages of Data Mining in Intrusion Detection, explained the sorts of Data Mining and Data Mining Process, and analyzed familiar methods and skills of Data Mining, and then evaluated them. For the representative traditional Association Rules Algorithm in Data Mining—Apriori Algorithm, this paper analyzed its deficiency, and gave an improved Association Rules Algorithm based on sequential Hash Table. This new algorithm was used in mining association rules in Database System. By these experiments, we can see the new algorithm not only reduced the time of mining Item-sets so as to enhance the efficiency of mining association rules, but also reduced the space for storing large candidate generations in Data Mining Processing so that can economize hardware resource. At last, to check the new algorithm's practical value, the paper structured a Database Intrusion Detection System base on the new algorithm to achieve of managing Database Safety. Preferable results were gained by experiments: the IDS can reach the need of the basic safety level in checking rate and accurate rate.