A Network User's Abnormal Behavior Detection Approach Based on Selective Collaborative Learning
Lu Yo · Chinese Journal of Computers · 2014
Machine learning techniques have been widely used in methods of network user's abnormal behavior detection.With the development of network,traditional detection methods cannot detect abnormal behavior accurately and quickly for their shortcomings such as cannot deal with unbalanced training data,huge demand for training data's label and so on.So this paper proposes a detection method based on selective collaborative learning.It uses improved EasyEnsemble algorithm to generate balanced training data.To improve the accuracy and reduce the cost of training, this method uses mixed perturbation algorithm to construct differentiated member classifiers and uses selective collaborative learning method to train them.Finally this method builds ensemble classifier according member classifiers' accuracy.Experiments showed this method can quickly and accurately detect abnormal behavior while reducing the demand for labeled training data.