论文部分内容阅读
开展二千分之一大比例尺土壤普查工作,有大量数据资料需要进行统计分析,以便获得各级样本特征的某些参数,其中尤以均值、标准离差最为重要。但在资料逐级汇总过程中,各项变量又要逐个输入计算器,重新运算这些参数。例如,宝山县有6899个耕层土壤样品的有机质、有效磷、有效钾的常规分析数据,加上分土种、土属、高程等等,则更显得繁浩,不仅工作量繁重,而且容易发生差错.为此,我们考虑是否可以通过各个样本的(?)、S_(n-1) 和 n 三个参数直接计算上一级的(?)、S_(n-1)、n 及 CV,这样工作量就可大大减少,还可以避免差错。由基本公式 S_(n-1)=((Σ(x-(?)~2)/(n-1))~(1/2),经推导换算获得如下公式:
Carrying out a one-thousandth of the scale of soil census work, there is a large amount of data required for statistical analysis in order to obtain certain characteristics of the sample parameters at all levels, of which the average, the standard deviation is most important. However, in the data level by level summary process, the variables have to enter the calculator one by one, to re-calculate these parameters. For example, there are 6,899 tillage soil samples in Baoshan County, which are routinely analyzed for organic matter, available phosphorus, and available potassium, together with soil types, soil types, elevations, etc., which are even more complex and not only heavy workloads but also prone to occur For this reason, we consider whether the upper-order (?), S_ (n-1), n, and CV can be calculated directly from the three parameters of (?), S_ (n-1) and n for each sample Workload can be greatly reduced, but also to avoid errors. The following formula is derived from the basic formula S_ (n-1) = ((Σ (x - (?) ~ 2) / (n-1)) ~