An Efficient Malware Classification Method Based on the AIFS-IDL and Multi-Feature Fusion

In recent years, the presence of malware has been growing exponentially, resulting in enormous demand for efficient classification methods. However, existing machine learning-based classifiers have high false positive rates and cannot effectively classify variants, packers, obfuscation. To address this shortcoming, paper proposes an deep method named AIFS-IDL (Atanassov Intuitionistic Fuzzy Sets-Integrated Deep Learning), which uses static features to malware. The proposed first extracts six types from disassembly byte files then fuses them solve single-feature problem traditional Next, Atanassov’s intuitionistic fuzzy set-based is used integrate result three learning models, namely, GRU (Temporal Convolutional Network), TCN CNN (Convolutional Neural Networks), improves accuracy generalizability model. verified by experiments results show that can improve compared Experiments were carried out on malicious code with algorithms ensemble algorithms. A variety comparative rate integrating multi-feature, multi-model aspects reach 99.92%. that, other methods, better identification ability.