Title:Benchmark Static API Call Datasets for Malware Family Classification

Abstract:Nowadays, malware and malware incidents are increasing daily, even with various antivirus systems and malware detection or classification methodologies. Machine learning techniques have been the main focus of the security experts to detect malware and determine their families. Many static, dynamic, and hybrid techniques have been presented for that purpose. In this study, the static analysis technique has been applied to malware samples to extract API calls, which is one of the most used features in machine/deep learning models as it represents the behavior of malware samples.
Since the rapid increase and continuous evolution of malware affect the detection capacity of antivirus scanners, recent and updated datasets of malicious software became necessary to overcome this drawback. This paper introduces two new datasets: One with 14,616 samples obtained and compiled from VirusShare and one with 9,795 samples from VirusSample. In addition, benchmark results based on static API calls of malware samples are presented using several machine and deep learning models on these datasets. We believe that these two datasets and benchmark results enable researchers to test and validate their methods and approaches in this field.