神经网络的Scilab

  • Z时代
  • 2024-01-10
  • 分类：问答

认识我试着运行这个例子，它采用神经网络工具箱的Scilab https://burubaxair.wordpress.com/2014/03/12/artificial-neural-networks-in-scilab/神经网络的Scilab

这是代码：

T = [ 
1 1 1 1 1 
0 0 1 0 0 
0 0 1 0 0 
0 0 1 0 0 
0 0 1 0 0 
0 0 1 0 0 
0 0 1 0 0 
]'; 
U = [ 
1 0 0 0 1 
1 0 0 0 1 
1 0 0 0 1 
1 0 0 0 1 
1 0 0 0 1 
1 0 0 0 1 
0 1 1 1 0 
]'; 
N = [35 10 2]; 
W = ann_FF_init(N); 
x = [1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
0, 1, 1, 1, 0]'; 
t_t = [1 0]'; 
t_u = [0 1]'; 
t = [t_t, t_u]; 
lp = [0.01, 1e-4]; 
epochs = 3000; 
W = ann_FF_Std_batch(x,t,N,W,lp,epochs); 
y = ann_FF_run(x,N,W) 
disp(y)

但是我收到一个错误：

-->exec('D:\Учёба\Задачи\Recognition.sce', -1) 
!--error 15 
Подматрица задана некорректно (Submatrix is incorrect). 
at line  37 of function ann_FF_grad_BP called by : 
at line  25 of function ann_FF_Std_batch called by : 
W = ann_FF_Std_batch(x,t,N,W,lp,epochs); 
at line  33 of exec file called by :  
exec('D:\Учёба\Задачи\Recognition.sce', -1)

一个错误可能在T和U矩阵中，但我不明白为什么。你能告诉我做错了什么吗？谢谢！

回答：

你让你的代码2级的错误：

  1. 你不应该混合测试和训练集。

  2. 测试输入必须是单列。

您的第一个错误是x = [ 1.... ]，因为它包含一个图像，而您在N中指定您有两个输出神经元。 正如例子说，你应该有x = [T,U];

你的第二个错误是给X作为测试ann_FF_run。该功能将测试输入作为单个列。但是因为你在x是5x7矩阵之前用x训练了你的NN。只需将其更改为列向量。

这里校正和注释的代码：

T = [... 
1 1 1 1 1 ... 
0 0 1 0 0 ... 
0 0 1 0 0 ... 
0 0 1 0 0 ... 
0 0 1 0 0 ... 
0 0 1 0 0 ... 
0 0 1 0 0 ... 
]'; 
U = [... 
1 0 0 0 1 ... 
1 0 0 0 1 ... 
1 0 0 0 1 ... 
1 0 0 0 1 ... 
1 0 0 0 1 ... 
1 0 0 0 1 ... 
0 1 1 1 0 ... 
]'; 
// setting the traing set of two image 
xtrain = [T,U]; 
// so each image as 35 pixels so N(1) is 35 
// and we have two images so N($) is 2 
N = [35 10 2]; 
// training the NN 
W = ann_FF_init(N); 
// The expected response for T : 1 for T, 0 for U 
t_t = [1 0]'; 
// The expected response for T : 1 for T, 0 for U 
t_u = [0 1]'; 
// the overall response 
t = [t_t, t_u]; 
// some parameters 
lp = [0.01, 1e-4]; 
epochs = 3000; 
// getting the weight of the trained NN 
W = ann_FF_Std_batch(xtrain,t,N,W,lp,epochs); 
// testing the traing set. 
y = ann_FF_run(xtrain,N,W) 
disp('Testing the traing set') 
disp(y) //should get something close to t ~ [1 0 ; 0 1] 
// testing a distord U 
xtest1 = matrix([1, 0, 0, 0, 1; 
1, 1, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
1, 0, 0, 0, 1; 
0, 1, 1, 1, 1]',-1,1); 
y = ann_FF_run(xtest1,N,W) 
disp('Testing a distored U') 
disp(y) //should get something close to t_u ~ [0 1] 
//testing something different from T and U. should get nothing 
xtest2 = matrix([1, 0, 0, 0, 1; 
1, 1, 0, 0, 1; 
1, 0, 1, 0, 1; 
1, 0, 1, 0, 1; 
0, 0, 1, 1, 1; 
0, 0, 1, 0, 1; 
0, 1, 1, 1, 1]',-1,1); 
y = ann_FF_run(xtest2,N,W) 
disp('Testing something neither T nor U') 
disp(y)

和SCILAB的控制台输出

Testing the traing set 
    0.8538757 0.1075397 
    0.1393287 0.8957439 
Testing a distored U 
    0.1078667 
    0.9007755 
Testing something neither T nor U 
    0.3433933 
    0.6306797

以上是 神经网络的Scilab 的全部内容， 来源链接： utcz.com/qa/265301.html

回到顶部