A comparative analysis of machine learning algorithms for faults detection in wireless sensor networks

Wireless sensor networks (WSNs) deployment experiences show that collected data is prone to be faulty. Faults are due to internal and external influences, such as calibration, low battery, environmental interference and sensor aging. In this paper, we focus on faults occurred due to low battery and calibration in WSNs. Machine learning algorithms have been successfully used to identify and classify various types of faults. In this paper, we evaluate and compare the performance of k-nearest neighbour, support vector machine (SVM), and Naive Bayes machine learning algorithms by using the real-world datasets to identify and classify the faults. We present here a comparative study of the above mentioned approaches on experimental datasets. The approach is validated using real data obtained from over one month of samples from motes deployed in an actual living lab. The results show that the k-nearest neighbour (kNN) algorithm obtained a better fault detection rate than other algorithms based on given performance metrics