Malware Detection based on API Calls Frequency
Vidhi Garg, Rajesh Kumar Yadav · 2019
Nowadays, one of the main dangers to computer system's security is the malware, a piece of software or a computer program that is designed to detriment and penetrate computers without the owner's permission. Traditional signature-based and anomaly-based malware detection approaches are still in use. However, the signature-based detection approach fails for new anonymous malware. In the anomaly-based detection, if the malicious activity behaves like a normal activity, the detection treats it as a normal one. Today's attackers are using various obfuscation techniques which have become a great challenge for the detectors to detect the malicious content with the traditional malware detection techniques. In this research, supervised learning algorithms are used to detect malware using the concept of API Calls usage frequency in a portable executable format. The experimental results provide the accuracy of 93% in distinguishing malware from benign files.