中文

Journal of Applied Sciences ›› 1995, Vol. 13 ›› Issue (2): 179-183.

• Articles • Previous Articles     Next Articles

ARTIFICIAL NEURAL NETWORKS AS APPLIED TO QUANTITATIVE SPECTROSCOPIC ANALYSIS

LI ZHILIANG1,2, LIU SHUQIONG1, SHI LEMING2, PAN ZHONGXIAO3, LOU MANTENG4, LI MENGLONG1, YU BANMEI5   

  1. 1. Hunan University;
    2. Institute of Chemical Metallurgy, Academia Science;
    3. Hunan University;Institute of Chemical Metallurgy, Academia Sinica;
    4. Changsha Institute of Technology;
    5. Guangzhou Institute of Anal
  • Received:1991-06-01 Revised:1994-07-26 Online:1995-06-30 Published:1995-06-30

PDF

73

Cited

2

Abstract: The artificial neural networks (ANN) method is utilized in the analytical calibration and quantitative resolution for spectroscopy. The simulated and practical data are proeessed by using the multilayered perceptrons model and backpropagation (BP) algorithm. The results show that is practicable to adopt artificial neural networks to calibrate and resolve the spectroscopie data and peaks.

Key words: multivariate spectroscopic analysis, chemometrics, artificial neural networks, backpropagution algorithm, calibration and resolution