Validation dataset for Probablistic Neural Network
3 views (last 30 days)
Show older comments
Hello. I am trying to apply Probabilistic Neural Network on a dataset. The command for Training on a dataset is: "net=newpnn(Train_Data,Train_targets,spread)" The problem I am facing is that it over trains on dataset. This can be prevented by using some of instances for Validation. But I do not know, how to define validation dataset before the training of PNN. Like we can define for patternnet and fitnet before their training: net.divideFcn = 'dividerand' I am facing the same problem with "newrb". Can anyone help me in defining validation dataset before training so that over training can be avoided.
0 Comments
Accepted Answer
Greg Heath
on 12 Dec 2015
NEWPNN is not used with a val set to prevent overtraining an overfit net.
For info on designing with NEWPNN search previous posts
SEARCH NEWSGROUP ANSWERS
newpnn greg 29 13
newpnn 47 29
Hope this helps.
Thank you for formally accepting my answer
Greg
2 Comments
Greg Heath
on 16 Dec 2015
As I have posted above, there are 76 posts for newpnn. You will have to search through them to find what you need.
Greg
More Answers (0)
See Also
Categories
Find more on Define Shallow Neural Network Architectures in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!