A Two-Layer Dimension Reduction and Two-Tier Classification Model for Anomaly-Based Intrusion Detection In IOT

Any device that can connect to the Internet and gather and transmit data is considered a "Internet of Things" device. As Internet connections have increased and new technologies like the Internet of Things (IoT) have been used, serious privacy and security concerns have emerged. Numerous, highly advanced incursions are introducing the IoT paradigm onto computer networks. Companies are increasing their research investments to detect these dangers more effectively. Institutions select intelligent techniques for testing and verification by comparing the greatest rates of accuracy. There has also been an increase in the usage of IoT in other industries, like health, in recent years. where technological researchers and developers are becoming more interested in IoT applications. Unfortunately, the privacy and security risks that result from IoT device energy limitations and scalability are the most obvious IoT challenge. Therefore, enhancing IoT security and privacy issues continue to be crucial issues in computer security. Using a two-layer dimension reduction and a two-tier classification module, we present a novel model for intrusion detection in this study with the aim of identifying risky behaviors like User to Root (U2R) and Remote to Local (R2L) attacks. The suggested model converts a high-dimensional dataset into a lower-dimensional dataset with fewer characteristics (LDA) using principal component analysis (PCA) and linear discriminate analysis from the dimension reduction module. To identify suspicious activity, we employ a two-tier classification module with Nave Bayes and a Certainty Factor form of K-Nearest Neighbor. Our model performs better than earlier models created to detect U2R and R2L assaults, according to tests on the NSL-KDD dataset. Not to mention, the testing findings showed that the PCA-LDA method performed better than other intrusion detection methods now employed in IoT technology.