• Linta Aniyan
• Ann Nita Netto

One of the most convenient solutions for detecting the failure in WSNs is the use of machine learning. Since WSNs have limited resources and are usually deployed in inaccessible and autonomous environments, each node in the network must be monitored to avoid adverse effects of faulty nodes on normal network operations. Fault detection in WSN is a challenging problem. The existing approaches for diagnosing faults in sensor networks leads to high complexity and low precision. A new fault detection mechanism based on Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) classifiers are proposed. In order to tune the classifier, feature selection method based on correlation is used so that the important or best features can be selected. Finally, the classification is done by the hybrid method that is combining the prediction score of both Support Vector Machines and K-Nearest Neighbor classifier. The addition of feature extraction will improve the fault detection performance in terms of accuracy and detection rate. So this method outperforms all other conventional methods.

Support Vector Machine (SVM); K-Nearest Neighbor (KNN); Detection Accuracy; False Positive Rate

"A Machine Learning Based Hybrid Approach for Fault Detection in WSN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 5, page no.198-201, May 2019, Available :http://www.jetir.org/papers/JETIRCV06039.pdf

"A Machine Learning Based Hybrid Approach for Fault Detection in WSN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 5, page no. pp198-201, May 2019, Available at : http://www.jetir.org/papers/JETIRCV06039.pdf