A HOST ANOMALY DETECTION METHOD BASED ON LDA MODEL
He Xi,Jiang Jianchun,Ding Liping,Wang Yongji,Liao Xiaofeng
DOI: https://doi.org/10.3969/j.issn.1000-386X.2012.08.001
2012-01-01
Abstract:The technique of intrusion detection based on sequence of host system call is a security detection technique mainly focusing on analysing the data set of host system call and further finding the intrusion.Its key technology relies on how to extract the characteristics of system call sequence more accurately and then followed by classification.In this paper,aiming at this,LDA(Latent Dirichlet Allocation) text mining model is introduced to build a new intrusion detection classification algorithm.In this method,topic characteristics of system call sequence are extracted using LDA model which the short sequence of system call is regarded by the method as word.Combined with the frequency characteristics of system calls,kNN(k-Nearest Neighbor) classification algorithm is used for anomaly detection.Experiment is evaluated on 1998 DAPRA data set,the result shows that the method improves the accuracy of intrusion detection,and reduces the false alarm rate.