Employing Methods Such as CNN2D and a Voting Classifier to Identify Ransomware Attacks Using Processor and Disk Usage Data

The research tackles the issue of ransomware detection, spotting the shortcomings of current strategies that rely upon system monitoring and information analysis. The objective is to create a dependable and effective detection mechanism for ransomware running on a virtual machine (VM). The data collection objectives precise processor and disk I/O activities for the complete virtual gadget from the host device. The studies is to develop an efficient detection model by means of making use of machine learning (ML), particularly a random forest (RF) classifier. This approach reduces monitoring burden and lessens the threat of data compromise by ransomware. The recommended method reveals robustness against fluctuations in user workloads, addressing a customary difficulty in ransomware detection. with the aid of eschewing incessant oversight of all techniques on the target pc, the version retains its adaptability to diverse user contexts. The task's efficacy is classed via diverse person workloads and 22 ransomware specimens. This venture gives a realistic and effective answer to the chronic ransomware hassle with a reliable detection mechanism. The project employs particular processor and disk I/O activities along gadget gaining knowledge of to reduce tracking overhead, improve detection pace, and keep adaptability to rising ransomware traces. This observe delivered huge upgrades, combining a Convolutional Neural network 2d (CNN2D) and an ensemble model with a voting classifier to enhance ransomware detection accuracy. The voting classifier, such as various machine learning classifiers, carried out an excellent ninety nine% accuracy in final predictions, illustrating the efficacy of version integration for enhanced detection.