Intelligent approach for detecting malicious PE files and URLs using machine learning

In the fast-growing era of technology, we have shifted from offline to online for our day-to-day personal or professional activities. As a vast amount of our data is available online; from individuals going about their daily lives to companies storing sensitive information, it is an easy target for cyber-criminals. These cybercrimes are carried out by various means such as corrupted PE files or via malicious URLs. There are various phishing sites or spam emails that send corrupted PE files or malicious URLs. In this paper, we propose a solution to this problem with a malware detection system that focuses on URLs and PE files. The system takes two types of inputs, URLs or PE files. The PE file once uploaded is processed to extract features (such as headers and section information) to help our ML-based classifiers to label it as malign or benign. For URL submissions, the system removed this lower-quality data (such as .com, HTTP) and pre-processed these noisy URLs to analyze for malicious intent. Through URL sanitization and PE file feature extraction, the system improves malware detection. As of now, the method that are proposed either work for PE file detection or URLS detection But this paper proposes a design that can detect both PE files and URLs.