Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Chinese Journal of Analytical Chemistry》 1994-09
Add to Favorite Get Latest Update

Study on the Multicomponent Determination by Artificial Neural Networks-Ultraviolet Spectrophotometry

Pan Zhongxiao,Wang Shuanhu, Chen Wei, Zhang Maosen (Departmet of Modern Chemistry, University of Science and Technology of China, Hefei 230026)  
The effect of various specifications and adjustable parameters on the training and prediction of artificial neural networks (ANN) in the determination of mixtures contaning four amino acids is studied. Experimental result indicates that a low learning rate(η) helps to get a good prediction, a high learning rate can make network to converge more quickly, but an overfit also emerges at the same time. Other network specifications and parameters, such as bias, momentum(α), gain(θ), number of hidden nodes, error limit, wavelength interval and transfer function, were also investigated and optimized. Considering both of run speed and prediction error of ANN, a set of ANN specifications and parameters for the purpose of calibrating multicomponent spectroscopic assay are proposed, They were tested by the experimental data, including one set of mixture containing amino acids and another set containing food pigments.
【Fund】: 国家自然科学基金
【CateGory Index】: O657.3
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
©2006 Tsinghua Tongfang Knowledge Network Technology Co., Ltd.(Beijing)(TTKN) All rights reserved