ARCHIVES

Abstract: Malicious assaults, malware, and ransomware families offer serious security challenges for cyber security, and they have the potential to cause catastrophic harm to computer systems, data centres, online, and mobile applications across a wide range of sectors and enterprises. Software (malware) has appeared and is growing in many formats and is becoming increasingly sophisticated. Criminals use them as a tool to infiltrate, steal or falsify information, causing huge damage to individuals, businesses and even threatening national security. It is a complex and varied threat that affects users globally, preventing them from accessing their system or data by locking the system's screen or encrypting and encrypting the users' files unless a ransom is paid. Traditional anti-ransomware technologies are unable to combat newly developed sophisticated assaults. As a result, cutting-edge approaches such as conventional and neural network-based designs can be very beneficial in the creation of unique ransomware solutions. In this project, propose a feature selection-based system for ransomware detection and prevention that uses deep learning methods, including neural network-based designs. We employed Multi-layer Perceptron classifiers on a sample of characteristics to classify malware. Then, to evaluate our proposed technique, we conducted all of the experiments on a single ransomware dataset. In terms of accuracy and precision ratings, the experimental findings show that MLP classifiers outperform other techniques.