建立神经网络输入参数过多怎么办

yuchaowangok

>  net=newff(p',t,[300,25,1],['tansig','tansig','purelim'],'trainrp','mse');
9 G  |% B2 l; N( J; A' Z7 b??? Cell contents reference from a non-cell array object.
* l9 Q. u* i4 Q% F
Error in ==> newff>new_5p1 at 199
2 p# M) r3 e3 h, j& h) t2 G  if (length(tf) < i) || all(isnan(tf{i}))

- q( t' [; E1 aError in ==> newff>create_network at 130
# ~- C  W) r, h4 C8 X- y  net = new_5p1(varargin{:});
6 e# {( s8 z' v% m- H- a# p! z
Error in ==> newff at 102
% {' q1 S: }: W& w9 D6 k  out1 = create_network(varargin{:});

yuchaowangok

我的这个网络输入参数有30个，但输入元素个数是300个

darker50

   一般就是降维，让参数变少呗！