Malware Combination Detection in Industrial Control Terminal Collaborative Response System Using Multiple Algorithms Based on Deep Learning

Because of its relevance and complexity in industry, the industrial control terminal collaborative response system is popular among app developers and users, as well as the main target of malevolent criminals. The rapid spread of malware has caused significant harm to users of the industrial control terminal collaborative response system. In this study, we build a model by combining dynamic and static feature detection, compiling test software output information using Fuzz technology, and training with a method that combines dynamic and static interest-based expert recommendation model (DSIERM) algorithm with Bi-LSTM algorithm. Modeling enables specific analysis and classification of test results. Based on this, the TaintBench suite and framework are introduced to compensate for the lack of relevant content for static taint analysis of applications in industrial control terminal collaborative response systems and to increase detection efficiency.