Fast Identification Method of Malicious Code Based on Extreme Learning Machine

A data balancing strategy based on conditional generative confrontation network and particle swarm algorithm can effectively improve the model's ability to detect malicious code. Nonetheless for some malicious code families with similar characteristics, there are still large errors in the model's detection results. This paper proposes a method for rapid identification of malicious code based on extreme learning machine(ELM). First, this paper introduce traditional single-hidden-layer feed forward neural networks (SLFNs) and analyze the weakness of the backpropagation algorithm. Then explaining the principle and advantages of the ELM algorithm. Finally, Introducing the detection method of malicious code based on the feature migration strategy and the ELM in detail. Experimental comparisons verify that the rapid detection model of malicious code proposed in this paper can effectively reduce the recognition time while ensuring high classification accuracy, and attain the requirements for real-time detection of malicious code.