A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Modeling water table depth using adaptive Neuro-Fuzzy Inference System
A comparative study on prediction of water table depth using Back Propagation Neural Network (BPNN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) is addressed in this paper. BPNN and ANFIS models are developed using precipitation, temperature, humidity, surface runoff and evapotranspiration loss to simulate water table depth fluctuations at Rengali, Odisha. Multiple linear regressions have been applied to observe the correlation between the dependent variable i.e. water table depth and the independent variables i.e. precipitation, temperature, humidity, surface runoff and evapotranspiration losses. Six different types of membership functions for ANFIS and three different transfer functions in case of BPNN are employed to compare the results of the models to predict water table depth. The model considering the physical parameters- precipitation, average temperature, evapotranspiration losses and humidity shows better performances on ANFIS as compared to BPNN. BPNN model using purelin transfer function produced best result with MAE testing 0.189701, RMSE testing 0.209842, and R2 testing 0.8512. ANFIS model with gauss2 membership function as input mf presented a superior ability to predict water table depth fluctuation over BPNN with MAE testing 0.080684, RMSE testing 0.089129, and R2 testing 0.9671. Taken as a whole, ANFIS is superior to BPNN for predicting water table depth.
Modeling water table depth using adaptive Neuro-Fuzzy Inference System
A comparative study on prediction of water table depth using Back Propagation Neural Network (BPNN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) is addressed in this paper. BPNN and ANFIS models are developed using precipitation, temperature, humidity, surface runoff and evapotranspiration loss to simulate water table depth fluctuations at Rengali, Odisha. Multiple linear regressions have been applied to observe the correlation between the dependent variable i.e. water table depth and the independent variables i.e. precipitation, temperature, humidity, surface runoff and evapotranspiration losses. Six different types of membership functions for ANFIS and three different transfer functions in case of BPNN are employed to compare the results of the models to predict water table depth. The model considering the physical parameters- precipitation, average temperature, evapotranspiration losses and humidity shows better performances on ANFIS as compared to BPNN. BPNN model using purelin transfer function produced best result with MAE testing 0.189701, RMSE testing 0.209842, and R2 testing 0.8512. ANFIS model with gauss2 membership function as input mf presented a superior ability to predict water table depth fluctuation over BPNN with MAE testing 0.080684, RMSE testing 0.089129, and R2 testing 0.9671. Taken as a whole, ANFIS is superior to BPNN for predicting water table depth.
Modeling water table depth using adaptive Neuro-Fuzzy Inference System
Das, Umesh Kumar (author) / Roy, Parthajit (author) / Ghose, Dillip Kumar (author)
ISH Journal of Hydraulic Engineering ; 25 ; 291-297
2019-09-02
7 pages
Article (Journal)
Electronic Resource
English
Accurate modeling of low-cost SiGe:C-HBTs using adaptive neuro-fuzzy inference system
| British Library Online Contents | 2005