XRF结合多元统计学对热敏纸的分类研究

Study on the Classification of Thermosensitive Paper by XRF Combined with Multivariate Statistical Analysis

关键词:  X射线荧光光谱  热敏纸  主成分分析  聚类分析

Key Words:X-ray fluorescence speetrometry  thermosensitive paper  principal component analysis  cluster analysis

摘要:为建立快速检验热敏纸的分析方法，本研究利用X射线荧光光谱仪（XRF）测定31种不同品牌、同一品牌不同系列的热敏纸样本，并结合多元统计学对实验结果进行主成分分析和系统聚类分析。根据样本所含元素种类及含量的不同，可有效区分31种热敏纸样本。当并类距离最小（为0）时，可将样品分为6类。验证的同时进行了判别分析，总体预测正确率为95.7%。该方法不破坏检材，重现性好，可为公安机关实际办案提供帮助。

Abstract:In order to establish an analytical method for rapid detection of thermal paper， 31 samples of different brands and different series of the same brand were determined by X-ray fluorescence spectrometry （XRF）， and the experimental results were analyzed by principal component analysis and cluster analysis combined with multivariate statistics. According to the kinds and contents of elements in the samples， 31 samples of thermal paper could be effectively distinguished. When the merging distance was the smallest， the samples could be divided into 6 types. Discriminant analysis was performed at the sametime of verification， and the overall prediction accuracy rate was 95.7%.This method did not damage the inspection materials and had good reproducibility， which could provide help for the public security sector in handling cases.