Skip to Main Content
A model of database anomalous detection is designed in this paper. The model can not only describe the users' behavioral profile more accurately, but also improve the accuracy of database anomalous detection. Based on the designed model, Apriori-kl algorithm, which combines the K-means clustering algorithm with the improved Apriori algorithm, is presented to mine users' behavior profile preferably so as to detect database anomaly more effectively and efficiently. Experimental results demonstrate that compared with the Apriori mining algorithm, Apriori-kl is superior in terms of time-consuming and detection accuracy.