Development of a Classification Model for Intrusion Attacks in Internet-Of-Things (IoT) Networks

Devices connected to the Internet can easily exchange information thanks to the Internet of Things (IoT), a networked system that functions via established protocols. IoT's decentralized architecture presents serious security, privacy, data integrity, and system stability issues despite its revolutionary potential. Although technological innovations like artificial intelligence, the Internet of Things, and big data have significantly raised people's quality of life, they have also increased the likelihood of increasingly complex and serious cyberattacks. By creating a machine learning model for the identification and categorization of intrusion threats in Internet of Things networks, this study seeks to address these issues. In particular, a hybrid strategy that combined Support Vector Machine (SVM) and fuzzy logic techniques was used to improve intrusion detection systems' efficacy in Internet of Things settings. We trained and tested the model using the NSL-KDD dataset from Kaggle. Key performance indicators, such as true positive rate, false positive rate, accuracy, F1 score, precision, and recall, were used to assess the suggested model's performance. With an accuracy of 99% on an imbalanced dataset and 98% on a balanced dataset, the results showed that the hybrid Fuzzy Logic-SVM model performed well across all data splits. These results demonstrate how the model may be used.