Volume 18, No. 6, 2021
Iot Enabled Dichotomous Regressive Ranking Decision Forest Node Classification For Efficient Data Transmission In Wireless Sensor Networks
L.Muthulakshmi and Dr.A.Banumathi
Abstract
An Internet of Things (IoT) permits several sensors that are connected to the Internet. The sensor nodes are a significant component of Wireless Sensor Networks assisted IoT networks to perform data acquisition for long-term monitoring. In this case, energy-efficiency is the most significant factor for long-term data acquisition to enhance the network lifetime. Therefore, developing a robust and energy-aware routing technique is a difficult task to expand the network lifetime. A novel IoT enabled Dichotomous Regressive Ranking Decision Forest Node Classification (IoT-DRRDFNC) technique is introduced for efficient data transmission in WSN. The IoT-DRRDFNC technique includes three processes namely data collection, classification, and data transmission. Initially, IoT devices are used for patient data collection at different locations. After that, sensor nodes are classified into two classes such as high-performance sensor nodes and less-performance sensor nodes by using the Dichotomous Regressive Ranking Decision Forest node Classifier. The sensor node has higher residual energy and minimum bandwidth consumption is classified as higher performance. In the IoT-DRRDFNC technique, a Dichotomous Regression tree is taken as a weak learner to classify the sensor nodes. Then the weak learner results are combined to make strong classification results by applying the ranking preferential voting scheme. After the classification, the only higher performance sensor node is taken for performing data transmission. Every sensor node selects the neighbouring sensor node with higher signal strength for minimizing the delay and packet loss rate during the data transmission in WSN. Experimental evaluation is carried out on factors such as energy consumption, packet delivery ratio, packet loss rate, and delay concerning several patient data packets and sensor nodes.
Pages: 5901-5917
Keywords: Wireless Sensor Networks, Internet of Things, Dichotomous Regressive Ranking Decision Forest node Classifier, Ranking preferential voting scheme, signal strength.