论文部分内容阅读
在林业科学研究(调查)中,研究对象的因子可以是定量的,如树高、树径;但也可以是定性的,如油茶的果型,立地条件的坡位、坡向。作为定性因子的坡位分上中下三等,果型为大小两级。显然各状态之间顺序是有意义的。对于这种顺序属性的因子,人们也试图以数据的大小依次表示其各种状态。虽然数量的大小反映了其顺序关系,但是却难于恰当地表示各状态间的差距,如上坡定数量为1,坡中应该定量为几?这种差距实在不好明确规定。而有些因子,诸如土壤母岩种类,树种别,品种别,则更难于用数值去分别表示它们了。在多元回归分析中,我们旨在研究某一因子(基准变量)与其它诸因子(说明变量)之间的定量关系。若说明变量中有定性的因子时,用常规的多元回归方法则不能处理,必须藉助于数量
In forestry scientific research (research), the research object factor can be quantitative, such as tree height, tree diameter; but also can be qualitative, such as the type of Camellia oleifera, site conditions slope position, slope. As a qualitative factor of the slope points on the lower third, fruit size of two levels. Obviously the order between the states is significant. For the factor of this sequential attribute, people also try to represent their various states in order of the size of the data. Although the magnitude of numbers reflects their order, it is difficult to properly represent the differences between the states. For example, if the number of slopes is 1 and the slope is to be quantified, what is the difference? The difference is not well defined. And some factors, such as soil rock type, tree species, species, it is more difficult to use numerical values to represent them separately. In multivariate regression analysis, we aimed to study the quantitative relationship between a certain factor (a benchmark variable) and other factors (explanatory variables). If there is a qualitative factor variables, using conventional multiple regression method can not be processed, we must use the number