Weighted Majority X-Means Ensemble Cluster Based Quadratic Discriminant Analysis for Resource Efficient Target Object Detection in WSN Using IOT
Energy efficiency is a considerable problem to be resolved during the process of target object detection in WSN because it determines the lifetime of the sensor network. Several research works are developed for target object detection in wireless network using different techniques. But, the object detection accuracy of existing techniques was poor. Besides, the amount of energy utilization was more. In order to overcome such limitations, Weighted Majority X-means Ensemble Clustering based Quadratic Discriminant Analysis (WMXEC-QDA) technique is proposed. Initially, the WMXEC technique considers numbers of sensor nodes that are arbitrarily positioned in wireless network and which are communicated using IoT. After that, WMXEC-QDA technique applies X-means Ensemble Clustering (XEC) algorithm where it generates ‘n’ number of weakX-means cluster result for each input sensor node in network. Then, XEC algorithm applies weights for each weak X-means cluster result. Subsequently, WMXEC-QDA technique designs a strong cluster by considering majority weights of all weak X-means cluster results with lower false positive rate. Finally, designed strong cluster in WMXEC-QDA technique groups each sensor node into consequent clusters with minimal amount of time complexity. Followed by, the WMXEC-QDA technique determines sensor node with higher residual energy as cluster head in order to effectively gather data about the target objects. After data gathering, cluster head forwards it to the sink node by means of the nearest cluster head. Then, sink node transmits the sensed data to the base station where it applies Quadratic Discriminant Analysis to preciselydetermine the target objects within the network. This assists for WMXEC-QDA technique to enhance the accuracy of target objects detection in WSN with lower time. The WMXEC-QDA technique conducts simulation work using metrics such as object detection accuracy, object detection time, false positive rate and energy usage with respect to a number of sensor nodes..