Prediction of ground water levels in uplands of coastal tropical riparian using PSO-SVM

Abstract

Wetlands are considered the most biologically diverse of all ecosystems. For the efficient management and proper development of a coastal tropical riparian, the behavior of ground water levels in uplands should be studied. Depending on the water levels in uplands, the surface water level on the nearby wetlands also changes, and sometimes changes the entire structure of wetlands. Computational Intelligence techniques such as Artificial Neural Network (ANN), Support Vector Machine (SVM) along with optimization techniques such as Particle Swarm Optimization (PSO) etc., can be used to develop a mathematical model for predicting the ground water levels in uplands. In the present paper, hybrid model PSO-SVM has been used with different kernel functions for the prediction of the ground water levels in the uplands of Padre Wetland in Surathkal, Mangalore, India. It was found that PSO-SVM tool with B-spline kernel function gives higher regression coefficients and also a stable model for the prediction of ground water levels in these uplands. © 2014 IJED.