How to define error weights in Neural Network?

2015 11 27
2 个回答
回答已采纳
更新时间：2015 12 16
6 次查看（30 天）

0 个投票

I want to minimize a mean squared weighted deviation in a neural network. The weights are different for each sample.
How can I specify this?
Is it perform(net,t,y,ew)?
ew = [a b c d ...] (weight for each sample)

请先登录，再回答此问题。

 采纳的回答

Greg Heath
Greg Heath 2015-12-16

0 个投票

The documentation commands
help train
doc train
yield
[NET,TR] = train(NET,X,T,Xi,Ai,EW)
Hope this helps.
Greg

更多回答（1 个）

Greg Heath
Greg Heath 2015-11-28
编辑：Greg Heath 2015-11-28

1 个投票

Yes.
However, I would scale the data so that the maximum weight is unity.
Hope this helps.
Thank you for formally accepting my answer
Greg

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Anderson
Anderson 2015-11-28
Hi Greg,
When I do perform(net,t,y,ew), the training process does not consider the weights. How can I include the weights in the training?
If I scale the weights (wide range interval), it will not loose information?

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

标签

Community Treasure Hunt

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

Start Hunting!

Translated by