Link Quality Prediction Based On Gradient Boosting

Abstract

To predict the consistency of the connections in the Wireless Sensor Network (WSN), which can aid in
the selection of links for the upper layer protocol. In this article we proposed a GBM (Gradient
Boosting Machine)-based prediction of the relation efficiency. This measures the relation parameters
(RSSI, LQI, SNR, and PRR) and trains the model to predict the future value of PRR based on real
experimental data. Experimental findings suggest that our proposed LQP-GBM prediction accuracy
indicates the efficacy of LQP-GBM as compared to CART AND LDA.