Downloads: 5 | Views: 83 | Weekly Hits: ⮙1 | Monthly Hits: ⮙4

Experimental Result Paper | Electrical & Electronics Engineering | India | Volume 13 Issue 1, January 2024

Innovative Approaches to Enhance Anomaly Detection in Wireless Sensor Networks

Naveen Kumar ML

Abstract: A crucial technology in many fields, wireless sensor networks (WSNs) provide for data collecting, monitoring, and management in contexts ranging from industrial settings to healthcare applications. Given the abundance of possible threats and weaknesses, the security of WSNs is still a major concern. In order to spot unexpected or hostile behavior inside these networks, anomaly detection is crucial. By utilizing cutting-edge machine learning techniques, this research proposes a novel strategy targeted at improving the effectiveness and accuracy of anomaly identification in WSNs. The suggested technique uses a combination of Random Forest, XGBoost, and K-Nearest Neighbours (KNN) classifiers inside an ensemble learning voting classifier framework to address the shortcomings of traditional anomaly detection methods. The two main goals of this integration are to reduce model complexity and improve classification accuracy. The work to obtain a more complete picture of the complex patterns available in WSN data by combining the capabilities of many classifiers. A crucial aspect of our approach lies in the utilization of the Infinite Feature Selection with Principal Component Analysis (PCA-IFS) technique. This hybrid feature selection method addresses the challenges posed by high-dimensional datasets, which are common in WSN applications. PCA-IFS not only identifies the most informative features for accurate anomaly detection but also addresses the curse of dimensionality. The fusion of PCA and IFS serves as a powerful mechanism to streamline complexity while ensuring the robustness of our approach.

Keywords: Wireless Sensor Networks, Anomaly Detection, Machine Learning, Ensemble Learning, Feature Selection

Edition: Volume 13 Issue 1, January 2024,

Pages: 733 - 738