Design And Implementation Of Trojans Behavior Feature Detection Algorithm Based On Deep Convolution Neural Network

As more and more industries has been united with the Internet,the "Internet plus" model has covered every aspect of life.However,along with the development with this model,some risks and challenges are also needs to be confronted,such as the technological innovation of Trojan horse and virus.Then this condition seriously affect the security of personal privacy and private information.According to the statistics of the national Internet emergency response center,the most important factor among these security problems is the Trojans,which means that it is of great significance to study the Trojans detection technology.Through the comparison of the existing Trojans detection technology,this paper determined that the main research target is how to detect the Trojans with dynamic behavior characteristics and divided it into two different parts.Based on the point of dynamic behavior characteristics,this research can be carried out as establishing the behavior characteristic library,establishing the data behavior characteristic and setting the data formatting rules of Gini coefficient.For another,based on the problem of classification algorithm,this research can be carried out as the design of deep convolutional neural network structure and the design of deep convolutional neural training method.Then with above two parts,this paper proposes the detection algorithm of Trojans behavior characteristics that based on the deep convolution neural network.At last,this paper has constructed the experimental environment,and tests the detection algorithm of Trojans behavior that based on the deep convolution neural network.Finally,the result of comparison experiment gives the advantages and disadvantages of the proposed detection algorithm.