AI Based Decision Tree Classifier for Detecting Anomalies in IoT Wireless Sensor Networks

This research study applies a Decision Tree Classifier to detect anomalies in IoT networks using the NSL-KDD collection of data, which is a used widely as a benchmark for intruder identification. Following categorical feature preprocessing with label encoding, a class-weight-adjusted Decision Tree was applied to handle class imbalance. The model achieved 92% accuracy and an F1 score of 0.89, showing very strong performance in distinguishing between normal and anomalous network behaviours. The results were further validated using performance indicators such as the ROC curve, classification matrix, and F1-score. This approach has great promise for real-time anomaly detection in IoT systems, enhancing network security with high accuracy and reliable threat identification. The novelty is in the integration of feature selection and hyperparameter optimization to improve the performance of the decision tree model, addressing the challenges of high-dimensional data and dynamic network conditions that have not been explored much in the existing literature.