Paper Info
Title
Anomaly detection in wireless sensor network using machine learning algorithm.
Abstract
Security in the Wireless Sensor Network(WSNs) is an essential and a challenging task. Anomaly detection is a key challenge to ensure the security in WSN. WSNs are vulnerable to various threats which may cause the node to get damaged and produce faulty measurements. The detection of such anomalous data is required to reduce false alarms. Machine learning algorithm based detection of anomalous data becomes popular now. Most of the current machine anomaly detection algorithms run in a stationary environment and require the entire training data to be kept in the node. In this paper, we formulate an Online Locally Weighted Projection Regression (OLWPR) for anomaly detection in Wireless Sensor Network. Linear Weighted Projection Regression methods are non parametric and the current predictions are performed by local functions that use only the subset of data. So, the computation complexity is low which is one of the requirements in Wireless Sensor Network. The dimensionality reduction in LWPR is done online by Principal Component Analysis (PCA) to handle the irrelevant and redundant data in the input data. After the prediction process, the dynamic threshold value is determined by a dynamic thresholding method to find the deviations of predicted value from the actual sensed value. OLWPR attains the detection rate of 86 percentage and very low error rate of only 16%.
Year
DOI
Venue
2020
10.1016/j.comcom.2020.01.005
Computer Communications
Keywords
Field
DocType
Wireless sensor networks,Anomaly detection,Machine learning,Linear weighted projection regression,LWPR
Anomaly detection,Dimensionality reduction,Computer science,Threshold limit value,Real-time computing,Artificial intelligence,Thresholding,Word error rate,Algorithm,Nonparametric statistics,Wireless sensor network,Machine learning,Principal component analysis
Journal
Volume
ISSN
Citations 
151
0140-3664
2
PageRank 
References 
Authors
0.37
0
2
Name
Order
Citations
PageRank
I. Gethzi Ahila Poornima120.37
Balasubramanian Paramasivan2353.53