Multilayer feedforward networks are universal approximators
Neural Networks
Fundamentals of neural networks: architectures, algorithms, and applications
Fundamentals of neural networks: architectures, algorithms, and applications
Neural Networks: A Comprehensive Foundation
Neural Networks: A Comprehensive Foundation
Advanced Methods in Neural Computing
Advanced Methods in Neural Computing
Using radial basis functions to approximate a function and its error bounds
IEEE Transactions on Neural Networks
Advances in Engineering Software
| Hi-index | 0.00 |
Bridge backwater data were collected for 92 different floods at 35 bridge sites in the Mississippi River basin in 1960s [Neely BL. Hydraulic performance of bridges, hydraulic efficiency of bridges-analysis of field data. Unpublished Report Conducted by US Geological Survey, June 30; 1966]. This major field data showed that the backwater computed both by the United States Geological Survey's method (USGS) and the United States Bureau of Public Roads' method (USBPR) averaged approximately 50% less than the measured backwater. Therefore, in the current work, a new bridge backwater formula based on the three different artificial neural network approaches (ANNs), namely FFBP (Feed-Forward Back Propagation), RBNN (Radial Basis Function-Based Neural Network), and GRNN (Generalized Regression Neural Networks) are proposed and compared with the methods mentioned above. The results showed that the FFBP produced slightly better estimations than those of the RBNN and these two was significantly superior to the GRNN, USGS and USBPR methods when applied to Neely's field data.