The design of database anomalous detection model based on user behavior profile mining

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.