Parallel-processed Hyperparameter Tuning for Higher Accuracy of Malware Detection

Many research works have focused on malware detection methods due to malware has become a major cybersecurity threat.Conventional malware detection methods often use pattern matching for executable code.In recent years, machine learning has been actively used to detect malware.However, it is difficult to improve the accuracy of malware detection models using machine learning effectively.In this study, we optimized the hyperparameters of machine learning-based malware detection models by parallel processing to improve the accuracy of unknown malware detection.In our experiments, we applied the proposed method to six machine learning algorithms, including deep neural networks (DNN) and random forests.We confirmed effectiveness of our proposed method than previous works in detecting malware on Windows and Android.