逐步回归分析与人工神经网络用于多组分稀土元素的HPLC同时测定  

Determination of Rare Earth Elements with HPLC Using Stepwise Regression Analysis and Artificial Neural Network

作　　者：邵学广[1,1] 陈宗海[1,1] 赵贵文[1,1] 

机构地区：[1]中国科学技术大学化学系

出　　处：《稀土》1998年第6期51-54,共4页Chinese Rare Earths

摘　　要：本文首先用逐步回归分析法对三组分稀土元素的重叠色谱数据进行筛选，提高数据的显著性，减少数据量，然后用ＢＰ模型的人工神经网络进行训练与预测，对重叠色谱峰中的各组分含量进行了定量分析。结果表明：采用逐步回归分析法对数据进行处理后，不仅提高了训练速度，而且预测结果也得到了明显的改善。In order to improve the quality and decrease the data points number of the input vector of an artificial neural network(ANN),the stepwise regression analysis(SRA) was used.A three component overlapping chromatogram was investigated by the SRA and ANN,and the results show that both the training speed and the predicted results can be significantly improved by the method.

关 键 词：逐步回归分析 ANN 重叠峰解析 稀土分析 HPLC 

分 类 号：O614.33[理学—无机化学]