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 A  ]: h- n. ^+ }??? Cell contents reference from a non-cell array object.
/ S9 i2 g9 X. f$ C% y3 [7 P% x4 G) N
Error in ==> newff>new_5p1 at 199
  if (length(tf) < i) || all(isnan(tf{i}))
4 u# `9 i5 M3 t* H2 r
( s4 f2 |7 R0 V* R! I6 KError in ==> newff>create_network at 130
  H- N& X9 {1 {! o7 w  net = new_5p1(varargin{:});
& h; B0 D5 F4 s: r7 n( ^! d
- _% b& j4 `) y: ^+ R; w/ N5 KError in ==> newff at 102
  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！