Optimized Decision Tree Classifier for Data Aggregation in Wireless Sensor Networks Using IoT Sensor Data

The Internet of Things (IoT) is a network that allows physical objects, sensors, appliances, and other items to communicate with each other without requiring human intervention.Wireless Sensor Networks (WSNs) are the main IoT components.The Internet of Things and WSNs have several significant and nonessential applications in practically every facet of contemporary life.The proposed study suggests a Decision Tree (DT)-based model that can be used to carry out data aggregation in an efficient and effective manner, utilizing the Intel Berkeley Research Lab dataset, a collection of 54 sensors.The methodology consists of three major steps of preprocessing, i.e., cleaning up the data, handling outliers, and label encoding, and feature optimization using Recursive Feature Elimination (RFE).The DT classifier is used to accomplish classification tasks.AUC-ROC, F1-score, recall, accuracy, and precision are used to evaluate measurements.It has been experimentally proven that the DT model exhibits the best classification accuracy of 97% and an AUC of 0.9928, exceeding the performance of other baseline models, including two-layer LSTM and Naive Bayes.The relative analysis validates the strength, interpretability and computability of the DT classifier, and thus it is an appropriate and viable solution to the WSN-based Smart IoT applications in terms of data aggregation.