Securing the Internet of Things: Using Machine Learning for Malware Detection with CIC-MalMem Dataset

The relevance of tackling security problems and assault risks in the Internet of Things (IoT) space is discussed in this study, focusing on malware detection. Malware is currently among the most important problems affecting internet users. The work emphasizes how weak security mechanisms in IoT devices leave them vulnerable to attacks like spyware, ransomware, and trojans, as well as to unauthorized access and data leaks. The importance of consumers exercising due caution, manufacturers prioritizing security, and creating industry standards are all emphasized. Machine learning (ML) approaches can analyze behaviours and traits in IoT networked gadgets to find potential malware or harmful activity. These algorithms can be taught to recognise and predict emerging threats as they arise, by being trained on massive datasets of well-known attack patterns. Several ML strategies as Support Vector Machine (SVM), Decision Tree (DT), Logistics Regression (LR) and Random Forest (RF) have been used for detecting the malware based on the CIC-MalMem-2022 dataset. Measures including accuracy, sensitivity, F1-score and precision have been utilised to evaluate these models' performance. The results show that the suggested approaches effectively detect network threats and improve IoT systems' capabilities. The analysis found that RF had the greatest accuracy rate of 98.45%, compared to LR, DT, and SVM, with accuracy rates of 96.83%, 97.67%, and 93%, respectively.