A Malware Detection Method Based on Genetic Algorithm Optimized CNN-SENet Network

In recent years, convolutional neural network (CNN) has achieved great success in the field of network security protection. With the popularization of smart terminals and the gradual increase of power grid informatization and digitization, the protection of power monitoring systems from various cybersecurity threads is a current scientific problem that needs to be solved urgently. To this end, this paper proposes a malware detection method based on genetic algorithm optimization of the CNN-SENet network, which firstly introduces the SENet attention mechanism into the convolutional neural network to enhance the spatial feature extraction capability of the model; then, the application programming interface (API) sequences corresponding to different software behaviors are processed by segmentation and de-duplication, which in turn leads to the sequence feature extraction through the CNN-SENet model; finally, genetic algorithm is used to optimize the hyperparameters of CNN-SENet network to reduce the computational overhead of CNN and to achieve the recognition and classification of different malware at the output layer. The examples under the public dataset containing 8 kinds of malware show that the proposed method is better than the traditional algorithmic model, and can accurately and efficiently achieve malware detection with strong generalization ability.