最优线性联想记忆网络方法在人发X射线荧光分析中的应用研究  

STUDY OF APPLICATION OF OLAM NEURAL NETWORKS IN HAIR XRF ANALYSIS

作　　者：袁学东[1] 吴丽萍[1] 龙先灌[2] 周厚全[1] 

机构地区：[1]四川大学物理系,成都610064 [2]四川大学原子核科学技术研究所,成都610064

出　　处：《四川大学学报（自然科学版）》2000年第4期558-562,共5页Journal of Sichuan University(Natural Science Edition)

摘　　要：作者针对X荧光分析中散射峰干扰和直线拟合精度差等问题 ,采用最优线性联想记忆 (OLAM )神经网络方法解谱 ,研究了无散射峰标准谱的产生和本底干扰问题 .通过用一维高斯分布随机数模拟本底计数产生无散射峰标准谱的方法 ,提高了分析精度 .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.

关 键 词：X射线荧光分析 OLAM神经网络 微量元素 人发 

分 类 号：Q581[生物学—生物化学] O657.34[理学—分析化学]