Chemical Research in Chinese Universities ›› 2001, Vol. 17 ›› Issue (2): 153-158.

• Articles • Previous Articles     Next Articles

Diagnostic Classification of Normal Persons and Cancer Patients by Using Neural Network Based on Trace Metal Contents in Serum Samples

ZHANG Zhuo-yong1, HONG Zhe2, ZHOU Hua-lan1, LIU Si-dong1   

  1. 1. Faculty of Chemistry, Northeast Normal University, Changchun 130024, P. R. China;
    2. Chemical Engineering Department, Dandong College, Dandong 118003, P. R. China
  • Received:2000-07-17 Online:2001-04-24 Published:2011-08-04
  • Supported by:

    Supported by Young Mainstay Teachers Foundation, Ministry of Education.

Abstract: Artificial neural network with the back-propagation(BP-ANN) approach was applied to the classification of normal persons and various cancer patients based on the elemental contents in serum samples. This method was verified by the cross-validation method. The effects of the net work parameters were investigated and the related problems were discussed. The samples of 72, 42, and 52 for lung, liver, and stomach cancer patients and normal persons, respectively, were used for the classification study. About 95% of the samples can be classified correctly. There fore, the method can be used as an auxiliary means of the diagnosis of cancer.

Key words: Artificial neural network, Classification, Trace element, Cancer