How do you make for get one/zero output?

2013 10 9
3 个回答
更新时间：2013 10 9
3 次查看（30 天）

0 个投票

the result not is iqual to the vector target because i get 0.5 in place of zero. What function i can add to my code for to get the output iqual to the target?
inputs =
Columns 1 through 11
6.3200 6.7800 7.5600 6.8400 9.0000 6.7200 6.5200 6.7400 7.0600 8.8400 6.2600
0 0 0 0 0 0 0 0 0 0 0
Columns 12 through 22
7.6400 8.1400 7.5200 8.0600 6.4800 6.5000 6.6600 7.2400 7.3200 5.1200 8.2400
0 0 0 0 0 0 0 1.0000 1.0000 1.0000 1.0000
Columns 23 through 32
6.7800 8.0200 8.0800 6.6600 7.7800 6.3400 8.3000 9.0000 7.2000 5.7800
1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
target =
Columns 1 through 19
0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 0
Columns 20 through 32
1 0 1 0 0 1 1 1 0 1 1 0 1
with the code: net=feedforwardnet(10); net.layers{1}.transferfcn='logsig'; net.layers{2}.transferfcn='logsig'; net.divideParam.trainRatio = 100/100; %net.divideParam.valRatio = 10/100; %net.divideParam.testRatio = 10/100; %net = configure(net,inputs,target); net=init(net); %net.iw{1,1} %net.b{1} %net.lw{2,1} net.trainparam.goal=0.001; net.trainparam.epochs=1000; net.trainFcn = 'trainlm'; %[net,tr]=train(net,inputs,target); net=train(net,inputs,target); output=sim(net,inputs) error=mse(target-output)
I get output: logit1
output =
Columns 1 through 11
0.5002 0.5002 0.5008 0.5002 0.9855 0.5002 0.5002 0.5002 0.5003 0.9101 0.5002
Columns 12 through 22
0.5009 0.5027 0.5007 0.5022 0.5002 0.5002 0.5002 0.5062 0.5085 0.5001 0.8629
Columns 23 through 32
0.5010 0.6980 0.7434 0.5006 0.5679 0.5002 0.9006 1.0000 0.5053 0.5001
error =
0.2121

请先登录，再回答此问题。

回答（3 个）

FRANK
FRANK 2013-10-9

0 个投票

My dear, I don't understand, could you say me in what place to the code write its?

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

dpb
dpb 2013-10-9
Any point at which you have the outputs of choice.
It simply demonstrates that you can create a logical vector of 0 and 1 by the comparison of the inputs to a suitably chosen reference level. What you do with that is dependent on your needs.
PS. It would help immeasurably if you would edit your initial question to properly format the code so it's not line-wrapped. Use the {}Code button or just prefix the first line of code with two blanks. Use the preview window to see what it looks like until get it right...

请先登录，再进行评论。

FRANK
FRANK 2013-10-9

0 个投票

net=feedforwardnet(10,'trainlm'); net.layers{1}.transferfcn='tansig'; net.layers{2}.transferfcn='tansig'; net.divideParam.trainRatio = 80/100; net.divideParam.valRatio = 15/100; net.divideParam.testRatio = 15/100; %net = configure(net,inputs,target); %net=init(net); net.iw{1,1} net.b{1} net.lw{2,1} net.trainparam.goal=0.001; net.trainparam.epochs=1000; %net.trainFcn = 'trainlm'; %[net,tr]=train(net,inputs,target); net=train(net,inputs,target); output=sim(net,inputs) error=mse(target-output)

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

dpb
dpb 2013-10-9
Still not formatted, unfortunately...and would be better if edit the original question, then remove this comment (and I'd follow up by cleaning up this one)

请先登录，再进行评论。

类别

在 帮助中心 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息

产品

标签

提问:

FRANK
FRANK
2013-10-9

评论：

dpb
dpb
2013-10-9

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by