Stage-discharge-sediment modelling using support vector machine
Author(s):
Manish Kumar and Pravendra Kumar
Abstract:
The sediment modelling are important aspects in planning and management of river dynamics. The stage and discharge play crucial role in sediment flow through a river channel. The present study carried out for stage-discharge-sediment modelling using support vector machine (SVM) and wavelet-based support vector machine (WSVM) for Adityapur site of Jharkhand, India. The best inputs variable was selected based on the gamma test. SVM and WSVM model were developed based on training (70% of total dataset) and testing (30% of total dataset). The developed models were analyzed through root mean square error (RMSE, g/l), Pearson correlation coefficient (PCC), Wilmott index (WI), line diagram and scatter plots. Based on line diagram and scatter plots, the results are under-predicted and over-predicted the SSC values. The results showed that the values of R2, RMSE, PCC and WI were found to be 0.3560 and 0.4078, 0.090 g/l and 0.077 g/l, 0.597 and 0.639, 0.744 and 0.767 for SVM and WSVM models, respectively during testing period. Therefore, WSVM model found to be superior and can be applied to predict daily SSC for Adityapur site.
How to cite this article:
Manish Kumar and Pravendra Kumar. Stage-discharge-sediment modelling using support vector machine. The Pharma Innovation Journal. 2021; 10(1S): 149-154.