论文部分内容阅读
李清同志在《元素背景值计算问题探讨》一文中,论述了柯氏检验、偏度与峰度检验在处理化探数据中的应用,无疑地具有实际意义,这里不再赘述.不过,就笔者所见,在该文的后半部份,还有值得商榷的地方.下面是笔者的一些浅见,提出来供讨论. 1.元素含量所遵循的分布问题该文指出了“同一批数据既符合算术正态分布,又符合对数正态分布时”的现象.原则上,随机数据所遵循的分布必须是图形一致或类似的分布.而正态分布与对数正态分布的图形在大多数场合下并不相似,只有在对数均值m较大、对数标准差s较小时的对数正态分布才趋近正态分布.此时,随机变量x_i的对数可表示为:
In his article “Discussion on the Calculation of Elemental Background Values”, Comrade Li Qing discussed the application of Korotkoff test, skewness and kurtosis test in the processing of geochemical data, which undoubtedly has practical significance and will not be repeated here I see, in the latter part of the article, there are debatable. The following is my humble opinion, put forward for discussion. 1. The distribution of elements to follow the issue pointed out that “the same batch of data both In line with the arithmetic normal distribution, but also with the logarithm normal distribution. ”In principle, the distribution of random data must follow the same pattern or similar distribution.But the distribution of normal distribution and logarithm normal distribution in the large Most cases are not similar, only logarithmic mean m is larger, the logarithmic standard deviation s when the logarithmic normal distribution tends to be normal distribution at this time, the random variable x_i logarithm can be expressed as: