Resolve of multicomponent mixtures using voltammetry and a hybrid artificial neural network method

Authors:
Shouxin Ren;Ling Gao
Affiliations:
Department of Chemistry, Inner Mongolia University, Huhhot, Inner Mongolia, China;Department of Chemistry, Inner Mongolia University, Huhhot, Inner Mongolia, China
Venue:
AICI'11 Proceedings of the Third international conference on Artificial intelligence and computational intelligence - Volume Part III
Year:
2011

Citing 1
Cited 0

Probabilistic neural networks

Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper suggests a novel method named DOSC-DF-GRNN, which is based on generalized regression neural network (GRNN) combined with direct orthogonal signal correction (DOSC) and data fusion (DF) to enhance the ability to extract characteristic information and improve the quality of the regression for the simultaneous simultaneous diffrential pulse voltammetric determination of Ni(II), Zn(II) and Co(II). In this case, the relative standard errors of prediction (RSEP) for total elements with DOSC-DF-GRNN, DOSCGRNN, DF-GRNN, GRNN and PLS were 9.70, 10.8, 11.5, 12.2 and 12.3 %, respectively. Experimental results showed the DOSC-DF-GRNN method was successful for diffrential pulse voltammetric determination even when there were severe overlaps of voltammograms existed and was the best among the five methods.