Default parameters in the RNN--WL model for network training

Input file names of hourly data required for RNN--WL model training and verification

Fig. 7. RNN--WL model training and verification at Station P8 of Fire Island Inlet.

Regional_Neural_Network
Page Navigation
11

W. Huang et al. / Ocean Engineering 30 (2003) 22752295

Default parameters in the RNN--WL model for network training

Optimized training method

Conjugated gradient training

Number of nodes in hidden layer

Training epoches

Note: When applying RNN--WL to different coastal regions, users may adjust these parameters to in

model training and verification.

Fig. 6. RNN--WL model training and verification at Station P2 of Shinnecock Inlet. The distance

between input NOAA station at Montauk and the inlet is about 60 km.

5. Model training, verification, and long-term predictions

5.1. Training and verification

In the model training and verification phases, two independent data sets of input

and output time series are required in the RNN--WL model. The first data set is

used for model training to determine the weight values of the interconnected network,

while the second data set is used for model verification by comparing the model

predictions with observations. Model parameters can be adjusted until model accu-

racy is satisfactory. After training and verification have been successfully completed,

Western Governors University

Privacy Statement
Press Release
Contact

Integrated Publishing, Inc. - A (SDVOSB) Service Disabled Veteran Owned Small Business