Open Access Open Access  Restricted Access Subscription Access
Open Access Open Access Open Access  Restricted Access Restricted Access Subscription Access

Evaluation of GRNN and RBF Model Performance for Groundwater Level Forecasting at Southwest Coast of India


Affiliations
1 Department of Civil Engineering, Sree Vidyanikethan Engineering College, Sree Sainath Nagar, Tirupati, A. Rangampet-517102, Andhra Pradesh, India
2 Department of Civil Engineering, Bearys Institute of Technology, Innoli, Boliyar Village, Near Mangalore University, Mangalore – 574153, Karnataka, India
     

   Subscribe/Renew Journal


An accurate and reliable forecast plays a vital role for proper planning and utilization of groundwater resources in a sustainable manner. In the present work, an investigation has been carried out in selective wells based on different land use/land cover in the micro watershed located southwest coast of India. The present study utilizes the Generalized Regression Neural Network (GRNN) for forecasting groundwater level (GWL) and compares its performance with that of the Feed Forward Back Propagation (FFBP) trained with Levenberg Marquartz (LM)] and Radial Basis Function (RBF). Weekly time series groundwater level data were used for span of three years (2004-2007). The comparative analysis of the obtained results showed that the GRNN and RBF have the superiority over the FFBP methods for forecasting groundwater level. On the basis of performance criteria (i.e lower RMSE and higher CE), GRNN yielded the better performance to RBF considering the models developed in the study.


Keywords

ANN, FFBP, GRNN, GWL, RBF.
User
Subscription Login to verify subscription
Notifications
Font Size

Abstract Views: 229

PDF Views: 2




  • Evaluation of GRNN and RBF Model Performance for Groundwater Level Forecasting at Southwest Coast of India

Abstract Views: 229  |  PDF Views: 2

Authors

Sreenivasulu Dandagala
Department of Civil Engineering, Sree Vidyanikethan Engineering College, Sree Sainath Nagar, Tirupati, A. Rangampet-517102, Andhra Pradesh, India
Nagaraj Gumageri
Department of Civil Engineering, Bearys Institute of Technology, Innoli, Boliyar Village, Near Mangalore University, Mangalore – 574153, Karnataka, India

Abstract


An accurate and reliable forecast plays a vital role for proper planning and utilization of groundwater resources in a sustainable manner. In the present work, an investigation has been carried out in selective wells based on different land use/land cover in the micro watershed located southwest coast of India. The present study utilizes the Generalized Regression Neural Network (GRNN) for forecasting groundwater level (GWL) and compares its performance with that of the Feed Forward Back Propagation (FFBP) trained with Levenberg Marquartz (LM)] and Radial Basis Function (RBF). Weekly time series groundwater level data were used for span of three years (2004-2007). The comparative analysis of the obtained results showed that the GRNN and RBF have the superiority over the FFBP methods for forecasting groundwater level. On the basis of performance criteria (i.e lower RMSE and higher CE), GRNN yielded the better performance to RBF considering the models developed in the study.


Keywords


ANN, FFBP, GRNN, GWL, RBF.