Malicious Code Detection Method Based on Multiple Features
Mingdi Xu, Hui Tong, Chaoyang Jin, Yu Wang · 2021
Malicious code detection has been considered as a major area for computer security. While a sharp increase in malicious code variants makes the accuracy and efficiency of the detection method reduced in a degree. To solve the problem, this paper proposes a multi-feature fusion method based on multiple N-value Opcode N-gram combined sequences and multi-scale gray image texture of malicious code. And then with the above fusion features, this paper uses RF and KNN machine learning algorithms to detect malicious code. At the same time, this paper takes accuracy, precision, recall, and f1 value as evaluation criteria to train and test massive malicious code samples. Finally, it verifies the effectiveness and accuracy of the malicious code detection method proposed in this paper through experimental results.