小麦中元素含量相关性分析的方法比较

Method comparison of correlation analysis of element content in wheat

目的建立小麦中元素含量相关性分析方法,并比较其相关性。研究对象为小麦中钡、钒、镉、锂、铝、锰、铅、铜、硒、铬、砷共11种元素。方法基于皮尔森相关系数法和偏相关系数法,建立小麦中11种元素的相关性分析,并对两种方法及结果进行比较。结果两种方法均能从数据中发现多种元素间的相关关系,并在运算复杂度、反映信息等方面各有特点:其中皮尔森相关系数法运算量较小,但反映的信息也较少;偏相关系数法的计算量和反映信息量都大于皮尔森相关系数。结论在当前的数据和软硬件条件下,建议采用偏相关系数进行小麦元素含量的相关性分析。

Objective To construct and compare the methods for correlation analysis of different elements in wheat,including barium,vanadium,cadmium,lithium,aluminum,manganese,lead,copper,selenium,chromium,and arsenic.Methods Pearson correlation coefficient and partial correlation coefficient were applied and compared to analyze the correlation among the 11 elements in wheat.Results Both of the methods could find the correlations among various contaminants from the data and had their own characteristics on computational complexity,information reflection and other aspects：Pearson correlation coefficient method need less calculation,but also reflected less information;partial correlation coefficient reflected more information but need more samples and computing resources.Under the current data and software and hardware conditions,the correlation analysis of the partial correlation coefficient was proposed.