Abstract:Because of the function of independent adaptation and study of NN and the best nonlinear speculating ability of fuzzy system, the adaptive network-based fuzzy inference system (ANFIS) which is composed of them absorbs their virtue. When applying it to the data analysis and model construction, the authors got the good nonlinear forecast by learning data, such as fluorescence spectrum. Fluorescence spectrometry is an important means of researching inside structure of molecule, which works easy, and features rapidness and high precision. So its forecast is more important. In the present paper, the authors utilize the spectra of the N2 to calculate, and prove that the means can forecast important spectrum information, the error is very small, less than 1.66 percent, which is conformed to meet the demand of the experiment. In short, the approach is workable.
[1] WEI Rong-hui, HUANG Yan-ping, LI Shan-shan, et al(魏荣慧, 黄燕萍,李珊珊,等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2005, 25(11): 1827. [2] XU Yi-zhuang,ZHAO Ying, XU Zhi, et al(徐怡庄, 赵 莹,徐 智,等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2005, 25(11): 1775. [3] LI Ying,ZHENG Rong-er,et al(李 颖,郑荣儿,等). Acta Laser Biology Sinica(激光生物学报), 2005, 14(5): 327. [4] Blank T B, Brown S D. Anal. Chim. Acta, 1993, 273; 277. [5] Lapedes A Farber. Technical Report LA-UR-87-2662, Los Alamos, NM: Los Alamos National Laboratory, 1987. [6] Weigend A B, et al. International Journal of Neural System,1990,1: 193. [7] Jang J S R. IEEE Transactions on Systems, Man, and Cybernetics, 1993, 23(3): 665. [8] Zadeh L A. Informat. Control, 1965,(8): 338.