STUDY OF APPLICATION OF OLAM NEURAL NETWORKS IN HAIR XRF ANALYSIS

The methods of OLAM neural network have been introduced into the analysis of the hair XRF analysis to overcome the problems such as the disturbance of scattering peaks and the poor accuracy of linear fit of data. Simulating the background spectrum with one dimension Gauss random numbers a method of producing the standard spectra without the scattering peaks is presented for the use of OLAM neural network techniques. The accuracy of measurement of hair XRF analysis is improved and the good measurement results are given.