论文部分内容阅读
逐步判别分析是一种多类判别的方法,它根据各变量的重要性大小,能逐个地把经F检验显著,即判别能力强的变量选入到别函数中,还能当引入新变量后,将原判别函数中的判别能力已减弱到使F检验不显著的变量剔出函数,直到再没有变量能引入函数,也没有已引入函数中的变量需要剔除时为止。计算判别系数把判别函数最终确定下来。将各个体观察值x_i(i=1、2、…m)代入判别函数,计算各y_i值,若y_k在y_i中为最大,即将该个体判为第K类。这种方法的优点是判别函数中的变量组合都是判别能力强的,从而提高了判别作用,使函数优化,但计算量大,须
Stepwise discriminant analysis is a multi-class discriminant method. According to the importance of each variable, it can be marked one by one by the F test, that is, the variable with strong discriminative ability is selected into another function, but also when a new variable is introduced. The discriminative power in the original discriminant function has been weakened until the variable that makes the F-test insignificant is discarded, until no more variables can be introduced into the function, and no variables that have been introduced into the function need to be discarded. The discriminant coefficient is calculated to finally determine the discriminant function. Substituting each body observation value x_i (i=1, 2,..., M) into a discriminant function, each y_i value is calculated. If y_k is the maximum in y_i, the individual is judged to be K-th class. The advantage of this method is that the variable combinations in the discriminant function are all strong in discriminative ability, which improves the discriminant function and optimizes the function, but the calculation volume is large.