Using Neural Network Technology In The EDXRF Analysis Basal Body Domino Offect Correction

In the development process of the EDXRF analysis, The Basal Body Domino Offect has been the objective factor which influences the EDXRF analysis sensitivity and accuracy and has restricted the level of analysis detection limit. So the EDXRF analysis basal body domino offect correction method has gone deep into much more.The traditional basal body domino offect correction methods mostly include: experiment correction method and mathmatics correction method.the methods can correct body body influence while satisfying the certain condition,exaltation analysis accuracy, But the usefulnesses of above-mentioned method all are depended on in consistency that used to be measured sample and mark kind type. Which is to say: the above-mentioned methods beseem the EDXRF analysis of simple sample.to the EDXRF analysis of the complicated sample,the basal body domino offect of now corrected meas to meet a challenge.Analyzed a research to provide a new science basic for the EDXRF from the all new angle.This text introduced the technique method of the nerve network,the point introduced SOFM network and RBF network, then described the EDXRF analytical basic characteristics,analyzed argument an usage the technical possibility of the nerve network,and used the technique the technique route of the method research.The complicated sample data mainly adopted a SOFM network,combined a characteristics of the complicated sample of Panzhihua minera area soil,got to the EDXRF analysis carried on an automatic classification a research and gived the EDXRF analytical classification regulation.then use a RBF network as to its won sample of type to carry on a basic body domino offect the dynamic state of the effect not the line correct a research,building up homologous mathematics modal.Pass to measure limited mark sample, attainding the rightness unkown the sample carries on an automatic classification;Shoot each element content in the same kind sample with its X in the meantime that the line records few rates not the line reflect to shoot a relation to come out through a mathematics model abstraction, building up not line a dynamic state an estimate model, coming to a to solve the purpose that the complicated sample EDXRF basal body domino offect correction.