Detection of Windows Executable Malware Files with Deep Learning

In today's internet age, malware emerges as a serious and growing threat in terms of information security. Therefore, detecting malware is extremely important in terms of preventing harm that malware may cause. In this study, by analyzing Windows Application Programming Interface (API) calls and the optional header sections of Windows executable files, it was tried to detect malware. A data set consisting of malware and benign executable files was created. In this study, 875 portable executable files were used, 283 of them are benign and 592 of them are malware. Each portable executable file in the data set is expressed in vectors by the taking into account Windows application programming interface calls and the optional header sections. Dimension reduction was made on feature vector. The reduced attributes were trained and tested by Deep Learning and detecting malware was achieved. At the end of the study, it was achieved 100% accuracy with Deep Learning