主成分分析在地质样品分类与浓度预测中的应用研究

Study on Classification and Elemental ConcentrationPrediction of Geological Samples by Principal Component Analysis

用主成分分析方法研究地质样品的Ｘ荧光光谱强度与浓度的关系，对未知样分类并预测样品浓度。对标准化后的数据计算各样品的主成分得分，根据得分分布图可快速分类样品。对训练样品作主成分回归分析，建立降维的主成分回归模型，用主元回归预测各组分浓度，效果好于多元回归分析方法。在标样较少的情况下，采用非线性组合增加维数的主元回归分析方法，比直接主元回归法的预测结果理想。

The relationship between concentration and Xray fluorescence intensity in geological samples was investigated by principal component analysis (PCA). The classification of these samples was performed by applying PCA to the standardized data and plotting the graph of their scores. Based on the plot of scores, a principal component regression model was built and applied to predict the concentration of major components in unknown geological samples. The results show that PCA method provides better results in predicting concentration of components in samples over the multi-component analysis method. If there are not enough standard samples, the prediction accuracy can be improved by combining simulative samples into the training set. The method can be applied to quantitative prediction of the element concentration in geological samples.\=\=