Neural Network Pattern Recognition - Network Training Issue,

2013 4 7
0 个回答
更新时间：2017 9 13
3 次查看（30 天）

0 个投票

Hi every one,
I am trying to build up a neural network for power system fault analysis ... In first step in training the network I have got many problems,
The major one is that training doesn't get start at all! it says PERFORMANCE GOAL MET or rarely GRADIENT MIN REACHED !
Does any one know whats wrong or how can I fix this? I am new with neural networks and I am using mainly the toolbox GUI and partially coding to gradually learn.
Any help is really appreciated, Bests, Pooria

4 个评论
显示 2更早的评论 隐藏 2更早的评论

papoo
papoo 2013-4-10
I am still working on this issue comparing my arrays of input with ToolBox examples. I guess there is some thing wrong with my data so iteration doesn't start as it meets the 1. Performance goal (Probably Least Square Error) 2. Gradient Limit. Hence it Do Not start the iterations. While I also tried changing the performance goal using the code: net.trainParam.goal = 1e-12;
But it didn't work.
Any helps or suggestions are really appreciated.
Greg Heath
Greg Heath 2013-4-10
I cannot help unless you post command line code and results using one of the MATLAB nndatasets.
help nndatasets
Greg
papoo
papoo 2013-4-10
Thank you Greg for your response. Fortunately I just solved it. There were a problem with my data arrays.
Thanks,
Greg Heath
Greg Heath 2017-9-13
Comment by Nelson Chow about 5 hours ago
(Moved from ANSWER block by Greg & lost his address
...SORRY !!!)
> Got the same problem. How did you solve this > problem and what problem with your data array?

请先登录，再进行评论。

请先登录，再回答此问题。

回答（0 个）

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:

papoo
papoo
2013-4-7

评论：

Greg Heath
Greg Heath
2017-9-13

Community Treasure Hunt

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

Start Hunting!

Translated by