wnbp.m

此程序是用于模式分类的在matlabhu环境下的原代码

nntwarn off
p=[0.4344 0.9950 0.6000 0.3865 0.3800 0.4345 0.2168 0.7180;
   0.4345 0.7000 0.5200 0.4350 0.3380 0.4705 0.0000 1.0000]
%0.4346	1.0000	0.6093	0.3862	0.3777	0.4347	0.2167	0.7179
%0.4346	0.7179	0.5222	0.4345	0.3377	0.4710	0.0000	1.0000]
t=[0 0 0 0 1 1 1 1;
   0 0 1 1 0 0 1 1;
   0 1 0 1 0 1 0 1]


%  Training parameters are:
%    TP(1) - Epochs between updating display, default = 25.
%    TP(2) - Maximum number of epochs to train, default = 1000.
%    TP(3) - Sum-squared error goal, default = 0.02.
%    TP(4) - Learning rate, 0.01.
%    TP(5) - Learning rate increase, default = 1.05.
%    TP(6) - Learning rate decrease, default = 0.7.
%    TP(7) - Momentum constant, default = 0.9.
%    TP(8) - Maximum error ratio, default = 1.04.
%  Missing parameters and NaN's are replaced with defaults.


tp=[200 8000 0.01 0.7 1.1 0.9 0.8 1.04];
[b1,w1,w2,b2,w3,b3,i,tr_hs,hs]=hsx3(20,10,'logsig','logsig','logsig',princomp(p),t,tp);