Abstract
An application of a Nonlinear Autoregressive Exogenous Neural Network (NARX) to predict total suspended sediment concentrations (SST) for a water body located in central Chile (Francia Creek, Valparaiso) is presented. Input data consisting of precipitation and stream flow time-series were fed to the developed NARX, for prediction of daily SST concentrations for a whole year. Sensitivity analysis was used for achieving the best NARX configuration that provided the best fit of simulated vs. measured data for year 2014. Parameters varied during sensitivity analysis were: number of nodes, number of iterations, feedback and forward delays, years of daily data used as training dataset. The resulting NARX is an open-loop net, consisting of a 12-node hidden layer, 100-iterations, using the Bayesian regularization backpropagation algorithm. SST concentrations predicted by the NARX net agreed successfully with measured SST concentrations (r = 0.73, r2 = 0.53, NSE = 0.18, PBIAS = −13.6%, Index of Agreement = 0.87).
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
