Malware Classification using LSTM-CF Framework for Disk Forensic Analysis

In the current era, digital forensic investigators need specialized tools to extract digital footprints from hard disks. The number of crimes is rising dramatically. This has resulted in an increased number of unresolved cybercrime cases including malicious software, hacking and cyberfraud. Disk forensics or disk investigation is a massive task. It takes weeks to collect the traces of the 1TB hard drive and analyze them. It is a challenging task to detect reliable evidence because of the worldwide use and advancement of digital communication technologies. By using automated tools, only predictable areas of the disk are often investigated, and as a result, latent evidence in the hidden area might be missed. In this paper, we present Chaithu's Framework (CF) that gives us a clear picture of the steps to follow and implement a machine learning algorithm which is able to predict whether the request is related to malware or not by analyzing the large world Microsoft Malware Prediction dataset collected from Kaggle. It consists of a total of 7.8 million data samples with 84 features collected in real time from Windows 7,8,9 and 10 systems. Using predictive machine learning algorithms, frauds can be detected automatically and autonomous actions can be taken to prevent them.