ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

STEPWISE REGRESSION BASED SUPPORT VECTOR MACHINE FOR STABLE AND CONSISTENT DATA DELIVERY IN WSN

Journal: International Journal of Computer Science and Mobile Applications IJCSMA (Vol.5, No. 10)

Publication Date:

Authors : ;

Page : 131-152

Keywords : Wireless sensor networks; Stepwise Regression technique; Support Vector Machine; Convex Hull;

Source : Downloadexternal Find it from : Google Scholarexternal

Abstract

Wireless Sensor Networks (WSNs) plays an important function with self-organized and structureless wireless networks for data delivery. With different routing protocols, efficient route path is selected by using the stable links. The establishment of stable network topology ensures communication effectiveness without any disconnection or disruption, for providing reliable data transmission. But, present network protocol modifies the route path considerably due to different interferences and environmental changes. Though, it obtains minimum data stability. Therefore, Stepwise Regression based Support Vector Machine (SR-SVM) technique is introduced to improve stable and consistent data delivery in WSN. Stepwise Regression technique and Support Vector Machine are the two process involved in proposed SR-SVM technique. At first, Stepwise Regression technique is used to examine the neighbouring node and select the target node. Based on neighbouring node, energy efficient neighbouring node is detected to attain a stable data delivery with minimum energy consumption in network. With the help of detected energy node, target node is selected for efficient wireless communication. After that, consistent data delivery is attained in network through link quality measure by using Support Vector Machine (SVM). Here, the link quality of the neighbouring nodes is considered to choose the optimal route by minimizing the distance function. In addition, convex hull of the two classes is used in geometric representation of SVM. Here, the number of hop between source and sink nodes are determined. This helps to improve link quality and network consistency with reduced data packet loss. The performance of SR-SVM technique is evaluated with parameters such as throughput, energy consumption, data loss rate, and average time. The experimental result shows that SR-SVM technique achieves higher throughput and better energy consumption when compared to state-of-the-art-works.

Last modified: 2017-11-01 20:22:06