Machine Learning Techniques to Detect and Classify Ransomware

The system mainly focuses on the malware where it is a type of malicious software called ransomware designed in such a way to block access to a computer system until a sum of money is paid. The automatic classification system which has high-performance, high accuracy and efficiency based on multi-feature selection fusion of machine learning is proposed in this project. The manual heuristic inspection of malware analysis is not considered to be effective and efficient when it is compared against the high spreading rate of malware which is a serious threat. Hence, using machine learning techniques the automated behavior based malware detection is considered to be a profound solution. The classifiers used in this project are K-Nearest Neighbors (KNN), Random Forests, and Decision Tree. Since, many antiviruses have heavy impact on the user system and sometimes they are not able to detect new emerging harmful threats like ransomware and on the other hand, light weight antiviruses are not so effective in detecting and preventing malwares. As the data security with no heavy impact on the user system is our main concern. Our model can be a cloud based malware classification model. Therefore, without installing any third party application on user system, we can test and predict whether any suspicious file is a malware or not.