新版Matlab中神经网络训练函数Newff的使用方法

青旅半醒 2022-05-18 05:21 567阅读 0赞

**一、**  **介绍新版****newff**

**Syntax**

· net = newff(P,T,\[S1 S2...S(N-l)\],\{TF1 TF2...TFNl\}, BTF,BLF,PF,IPF,OPF,DDF)

**Description**

newff(P,T,\[S1 S2...S(N-l)\],\{TF1 TF2...TFNl\}, BTF,BLF,PF,IPF,OPF,DDF) takes several arguments

<table style="width:518.15pt;"> 
 <tbody> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> P </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> R x Q1 matrix of Q1 sample R-element input vectors </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> T </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> SN x Q2 matrix of Q2 sample SN-element target vectors </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> Si </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Size of ith layer, for N-1 layers, default = [ ]. (Output layer size SN is determined&nbsp;from T.) </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> TFi </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Transfer function of ith layer. (Default = 'tansig' for hidden layers and 'purelin' for output layer.) </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> BTF </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Backpropagation network training function (default = 'trainlm') </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> BLF </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Backpropagation weight/bias learning function (default = 'learngdm') </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> IPF </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Row cell array of input processing&nbsp;functions. (Default = {'fixunknowns','removeconstantrows','mapminmax'}) </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> OPF </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Row cell array of output processing functions. (Default = {'removeconstantrows','mapminmax'}) </td> 
 </tr> 
 <tr> 
 <td style="background-color:#ccffff;border-color:#808080;"> DDF </td> 
 <td style="background-color:#ccffff;border-color:#808080;"> Data divison function (default = 'dividerand') </td> 
 </tr> 
 </tbody> 
</table>

**Examples**

Here is a problem consisting of inputs P and targets T to be solved with a network.

· P = \[0 1 2 3 4 5 6 7 8 9 10\];T = \[0 1 2 3 4 3 2 1 2 3 4\];

Here a network is created with one hidden layer of five neurons.

· net = newff(P,T,5);

The network is simulated and its output plotted against the targets.

· Y = sim(net,P);plot(P,T,P,Y,'o')

The network is trained for 50 epochs. Again the network's output is plotted.

· net.trainParam.epochs = 50;net = train(net,P,T);Y = sim(net,P);plot(P,T,P,Y,'o')

**二、**  **新版****newff****与旧版****newff****调用语法对比**

**Example1**

**比如输入****input****（****6\*1000****），输出****output****为（****4\*1000****），那么**

旧版定义：net=newff(minmax(input),\[14,4\],\{'tansig','purelin'\},'trainlm');

新版定义：net=newff(input,output,14,\{'tansig','purelin'\},'trainlm');

**Example2**

**比如输入****input****（****6\*1000****），输出****output****为（****4\*1000****），那么**

旧版定义：net=newff(minmax(input),\[49,14,4\],\{'tansig','tansig','tansig'\},'traingdx');

新版定义：net=newff(input,output, \[49,14\], \{'tansig','tansig','tansig'\},'traingdx');

**三、**  **旧版****newff****使用方法在新版本中使用**

**提示：旧版本定义的****newff****虽也能在新版本中使用，但会有警告，警告如下：**

Warning: NEWFF used in an obsolete way.  
> In obs\_use at 18  
In newff>create\_network at 127  
In newff at 102  
See help for NEWFF to update calls to the new argument list.

**四、**  **新版****newff****与旧版****newff****使用的训练效果对比**

**旧版本：**旧用法训练次数多，但精度高

**新版本：**新用法训练次数少，但精度可能达不到要求

**造成上述原因是：**

程序里面的权值、阈值的初始值是随机赋值的，所以每次运行的结果都会不一样，有好有坏。  
你可以把预测效果不错的网络的权值和阈值作为初始值。  
具体可以查看net.iw\{1,1\}、net.lw\{2,1\}、net.b\{1\}、net.b\{2\}的值。

**现在给一个完整的例子**

**%%** **清空环境变量**

**clc**

**clear**

**%%** **训练数据预测数据**

**data=importdata('test.txt');**

**%****从****1****到****768****间随机排序**