A Malicious Code Detection Strategy Based on Feature Fusion
Liang Wu · 2024
Due to its weak characteristics, the general malware software detection technology has the weakness of inaccurate detection and inefficiency. Therefore, a malicious application or software detection mechanism is designed based on feature fusion. The detection mechanism based on OPC X-gram and malicious applications or software is improved, and the malicious applications or software detection mechanism based on OPC X-gram with multi-X value combination can mine the expressive logic sequence of malicious applications or software and improve the detection ability of malicious application or software. The potential feature representation mechanism of OPC X-gram is designed, and the decompiling component is used to get the source of the procedure to be tested. Then the source parts to be analyzed is extracted. Combined with the OPC X-gram sequences with multiple X values, the recognizable sequences are screened out by using the content feature recognition method, and the malicious application or software classification training is carried out on the OPC X-gram logic sequences with different X values by using KNN and RF classification algorithms, and finally the OPC X-gram will be used. The inferring sequence will be f made up by logic analysis again. By the experiment analysis, the proposed multi-X value OPC X-gram approach is better than the bin-file X-gram sequence logic and the single-X value OPC X-gram in terms of the accuracy of the potential threat application or software classification.