基于主成分分析的中药色谱指纹图谱多维多息特征数据挖掘方法研究

Principal component analysis based on the multi-dimensional information characteristics in TCM fingerprints

目的对评价中药色谱指纹图谱多维多息的F和I等37个特征参数进行数据挖掘,为用计算机解析与评价图谱、建立标准的指纹图谱提供理论参考和实践探索。方法运用统计软件SPSS对10批次不同产地的当归指纹图谱的37个多维多息特征参数进行主成分分析。结果运用药学专业知识对分析过程产生的4个主成分进行命名,计算主成分得分,其中S10批次样本得分最高为1.52。结论综合主成分得分排名可作为评价指纹图谱的模型,从对各主成分的命名中发现了4个主成分能够反映中药色谱指纹图谱信息的规律,证实了主成分分析能达到降维目的,使繁多的求解目标简化,可用于中药指纹图谱的数据挖掘。

Objective To data mine the 37 parameters such as F and I of the multi-dimensional information characteristics of traditional Chinese medicine, and to supply theoretical reference for practical research. Methods According to the flow of data mining, the principal component analysis of the 37 parameters of multi-dimensional information characteristics of 10 kinds of traditional Chinese medicine was done with SPSS, which were CHINESE ANGELICA from 10 places. Results The principal components were named applying the academic knowledge, and the principal components were calculated. Sample 10 scored higher than the other 9 samples, which was 1.52. Conclusion The rank of scoring of principal component analysis can be used as a module to evaluate the TCM fingerprints. Principal component analysis can be used to decrease the dimension and can play an important role in the data mining of the traditional Chinese medicine Chromatographic fingerprints. Principal component analysis can be used to do data mining for the fingerprint of traditional Chinese medicine.