化工冶金研究中线性图的评价

Appraisal of linear plots in chemical engineering and metallurgy

分析了化工冶金文献中使用最广泛的几种线性关系确定法的固有优缺点 ,认为 t分布法是确定线性关系参数最科学的方法 .该法结合了线性相关系数 r,置信度α,实验点数 N以及临界线性相关系数 r临 .并使用这种方法检验了文献中声称为线性关系的 89组数据 .结果表明 ,在 95 %的置信度下 ,受检的数据所构成的线性关系中有 1 5组不合格 ,占检验总数的 1 9% .这些不合格的线性关系 ,主要是由实验点数小于

This paper analyses the advantages and disadvantages of several methods of established linear relations used widely in research of chemical engineering and meatllurgy.It concludes that the t -distribution method is most scientific for defining linear correlation parameters in that it operates in an integrated way with the linear correlating coefficient' r ',correlating coefidence level ' a ', ' N ',the number of experimental points,and ' r c',the critical linear correlating coefficient.The paper further addresses how to use this method to examine the declared linear relationships in literatures.The results have shown that 15 groups are not statistically significant at a 95 % confident level of the 81 ones,taking 19 % of the total events.Majority of the 15 groups are composed of the data groups their numbers are less than 5.