Revolutionizing Connectivity and Data Processing in IoT Ecosystems through Edge Computing

This work, which explores the performance of the existing five IoT sensor classification algorithms and develops five more algorithms, intends to later compare the obtained classification accuracies. In this evaluation, the algorithms included are Decision Tree, Random Forest, K-Nearest Neighbors, Support Vector Machine, and Naive Bayes, while the proposed ones are called EdgeNet, IoTBoost, EdgeSVM, DeepEdge, and EdgeFusionNet. The performances are evaluated using the key measures of precision and processing time. Results show that the algorithms under study are usually more accurate than well-known ones, with EdgeFusionNet reaching 94% accuracy which is the highest accuracy. In the selection of algorithms, existing ones show different performance but the Decision Tree and Naive Bayes demonstrate relatively faster processing times compared to the Random Forest and the Support Vector Machine. Yet, the proposed EdgeSVM algorithm shows almost the fastest processing time which exceeds Zokfl2 and other algorithms onference. On top of that, extra measures comprise memory usage, energy efficiency, scalability, model size, and robustness are evaluated for the suggested algorithms. Thus, the results may be inferred that the suggested algorithms provide the advanced solutions for the IoT sensor data classification, that have higher accuracy and comparable or better time complexity in comparison with the existing ones. These outcomes give useful information for researchers as well as technicians that using IoT techniques for the sake of various applications purposes.