AI- BASED DETECTION AND PREVENTION OF WANNACRY RANSOMWARE ON WINDOWS SYSTEM

Ransomware attacks have become a big problem of cybersecurity over the recent years, and WannaCry remains to be one of the most catastrophic cases in history. This study introduces the idea of designing an Artificial Intelligence (Al)-centred system that will detect and prevent Windows-based WannaCry ransomware. The overall goal of the proposed project is to develop a smart system that will be able to detect ransomware behaviour without any human intervention through the use of machine learning, and reduce the chances of having files encrypted and impaired systems. The system uses a controlled learning process by a Random Forest Classifier which is being taught on a restricted dataset called Final_Dataset_without_duplicate.csv, retrieved at Zenodo. The dataset includes 2275 samples and 77 attributes which are a combination of benign windows files and WannaCry ransomware variants. Pretreatments consisted of the removal of irrelevant columns (e.g., This project) based on the undesirability of the column), which was aimed at the development of an Artificial Intelligence (Al)-based system that detected and prevented WannaCry ransomware on the Windows operating system. The primary goal of the project is to develop a smart framework that would be able to recognise ransomware subversion automatically with the help of machine learning to minimise the risks of file encryptions and compromise of the system. The system uses a form of controlled learning strategy using a Random Forest Classifier, which is given the Final_Dataset_without_duplicate.csv dataset acquired in Zenodo. The data itself has 2,275 samples and 77 attributes, consisting of both benign Windows files, and WiCry variants of ransomware. This is due to the fact that cybersecurity attacks that are ransomware related have become one of the most crucial cyber security threats of the last few years, and WannaCry is by far the most devastating example to date. The study introduces the creation of an Artificial Intelligence (Al)-driven tool that is capable of identifying and preventing WannaCry ransomware on the Windows operating systems. The primary aim of this project is to develop a smart system that will be able to detect ransomware activities automatically by applying the machine-learning approach, which will reduce the likelihood of file encoder and system vulnerability. The system uses the supervised learning method using a Random Forest Classifier on a filtered dataset Final_Dataset without duplicate.csv which was acquired on Zenodo. There are 2275 samples and 77 attributes in the dataset that include both benign windows files and the types of WannaCry ransomwares. Keywords: Artificial Intelligence, Machine Learning, Ransomware Detection, WannaCry, Cybersecurity, Windows Systems