Detecting Malevolent Software with Machine Learning Models

Abstract

The ceaseless movement of malware positions a basic challenge in cybersecurity, changing to mechanical headways in appear despise toward of executed security measures. This paper presents an inventive approach to upgrade the disclosure of cluttered malware through the integration of machine learning (ML). Utilizing a real-world dataset of otherworldly malware sorts such as spyware, ransomware, and trojan steeds, our consider addresses the advancing challenges of cybersecurity. In this consider, we assess the execution of ML calculations for tangled malware disclosure utilizing the CIC-MalMem-2022 dataset. Our examination envelops parallel and multi-class classification errands underneath differing exploratory conditions, counting rate parts and 10-fold cross-validation. The studied calculations solidify Subjective Tree (RT), Scattered Timberland (RF), J-48 (C4.5), Unsophisticated Bayes (NB), and XGBoost. Test comes nearly layout the ampleness of RF, J-48, and XGBoost in wrapping up tall accuracy rates over unmistakable classification assignments. NB also appears up competitive execution but faces challenges in taking care of imbalanced datasets and multi-class classification. Our divulgences highlight the significance of utilizing progressed ML strategies for moving forward cluttered malware disclosure capabilities and give basic experiences for cybersecurity specialists and analysts. Future inquire around introduction connect finetuning show up hyperparameters, investigating gathering learning approaches, and creating assessment to orchestrated datasets and real-world scenarios.