In this project, Bottleneck Stacked Autoencoder, a
deep learning method, is used to model the normal program
(application) behavior and to detect abnormal behavior in terms
of the system call usage. The method is evaluated by applying the
proposed approach to an example application in which malicious
code are embedded. The evaluation results show its effectiveness
in terms of the detection accuracy and false positive rates, which
are even comparable to Gaussian Mixture Model.
more here.....http://courses.engr.illinois.edu/ece544na/fa2013/writeups/ECE544NA_final_report_mkyoon.pdf
deep learning method, is used to model the normal program
(application) behavior and to detect abnormal behavior in terms
of the system call usage. The method is evaluated by applying the
proposed approach to an example application in which malicious
code are embedded. The evaluation results show its effectiveness
in terms of the detection accuracy and false positive rates, which
are even comparable to Gaussian Mixture Model.
more here.....http://courses.engr.illinois.edu/ece544na/fa2013/writeups/ECE544NA_final_report_mkyoon.pdf