Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

DBN+BPNN for fault classification #182

Open
SJHMAX opened this issue Jul 5, 2018 · 8 comments
Open

DBN+BPNN for fault classification #182

SJHMAX opened this issue Jul 5, 2018 · 8 comments

Comments

@SJHMAX
Copy link

SJHMAX commented Jul 5, 2018

I use DBN for the classification of bearing fault data sets at Case Western Reserve University. The data set includes 10 failures, but the final classification result is always a category. When you modify various parameters of the network in nnsetup.m, the result will change accordingly. For example, when you change nn.learningrate to 1, the classification will change from 1 to 4. Many people have encountered this problem, but No suitable solution was found. Sincere hope that someone can help me

@joey0111
Copy link

joey0111 commented Jul 5, 2018

You can try to change the initial weight of the RBM and increase the number of training.

@SJHMAX
Copy link
Author

SJHMAX commented Jul 5, 2018

@joey0111
thanks for your help.
i tried to set the initial weight to a random number between 0 and 1, and increased the number of training.The result is still the same as before.

@joey0111
Copy link

joey0111 commented Jul 5, 2018

A random number between 0 and 1 is not enough. You can try multiplying all the initial weights by 0.01 or a smaller number. Also, have you normalized the input data?

@SJHMAX
Copy link
Author

SJHMAX commented Jul 5, 2018

@joey0111
thanks....but it si still not working...and All data has been normalized.
The data source is the open bearing fault data set of Case Western Reserve University, USA, which is the original time domain signal. The training samples include 10 types of faults, each of which has 600 training samples, a total of 6000 groups, and 128 data points per group (all normalized).

@Passion-long
Copy link

Have you solved this problem? It seems that I have come across the same question.

@Yi0377
Copy link

Yi0377 commented Apr 26, 2019 via email

@Passion-long
Copy link

Passion-long commented Apr 28, 2019 via email

@SJHMAX
Copy link
Author

SJHMAX commented May 20, 2019

I hava tried to use matlab2018b,but the result doesn't change at all.Is it possible that the code has a problem?

------------------ 原始邮件 ------------------ 发件人: "Yi Jing"notifications@github.com; 发送时间: 2019年4月26日(星期五) 下午3:44 收件人: "rasmusbergpalm/DeepLearnToolbox"DeepLearnToolbox@noreply.github.com; 抄送: "王云龙"1922631820@qq.com; "Comment"comment@noreply.github.com; 主题: Re: [rasmusbergpalm/DeepLearnToolbox] DBN+BPNN for faultclassification (#182) i use the new matlab2018b。 | | Eric | | lq283528870@163.com | 签名由网易邮箱大师定制 On 4/26/2019 15:07,Yunlong Wangnotifications@github.com wrote: Have you solved this problem? It seems that I have come across the same question. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

这个工具箱是没有问题的,就是不好调试,因为需要人工进行。

我用自己的数据,总共十类,用这个工具箱可以达到95的准确率。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants