For malicious purposes, attackers hide malware in the software used by their victims. New malware is continuously being shared on the Internet, which differs both in terms of the type of malware and method of damage. When new malware is discovered, it is possible to check whether there has been similar malware in the past and to use the old malware to counteract the new malware; however, it is difficult to check the maliciousness and similarity of all software. Thus, deep learning technology can be used to efficiently detect and classify malware. This study improves this technology’s accuracy by converting static features, which are binary data, into images and by converting time-series data, such as API call sequences, which are dynamic data with different lengths for each datum, into data with fixed lengths. We propose a system that combines AI-based malware detection and classification systems trained on both static and dynamic features. The experimental results showed a detection accuracy of 99.34%, a classification accuracy of 95.1%, and a prediction speed of approximately 0.1 s.
When using dynamic analysis data to detect malware, time-series data such as API call sequences are used to determine malicious activity through deep learning models such as recurrent neural networks (RNN). However, in API call sequences, APIs are called differently when different programs are executed. To use these data as input for deep learning, preprocessing is performed to unify the size of the data by adding dummy zeros to the data using the zero-padding technique. However, when the standard deviation of the size is significant, the amount of dummy data added increases, making it difficult for the deep learning model to reflect the characteristics of the data. Therefore, this paper proposes a preprocessing technique using term frequency–inverse document frequency (TF-IDF) and a sliding window algorithm. We trained the long short-term memory (LSTM) model on the data with the proposed preprocessing, and the results, with an accuracy of 95.94%, a recall of 97.32%, a precision of 95.71%, and an F1-score of 96.5%, showed that the proposed preprocessing technique is effective.