A***A
发帖数: 98 | 1 I'm not sure what is meant by your examples, but you may check out a
permutation based method that looks explicitly at the distance between the two
data vectors.
Here is the citation;
Mielke, Berry & Johnson (1976) Commun. Statist. - Theor. Meth., A(5), 14, pp.
1409 - 1424. | f******a
发帖数: 9 | 2 Fisher's linear discriminant
J(w) = (m2-m1)*(m2-m1)/(std1*std1+std2*std2) [1]
Reference:
[1] Biship M. Christopher, (1995), "Neural Network for Pattern Recognition",
Clarendon Press Oxford, ISBN: 0198538642, pp. 107
dispersion |
|
