get neural network performance outputs

Hi

I use matlab nnstart command to create and train MLP neural network and scripting them at final by its advanced script maker that have not any option for performance reporting is there any way to get neural network performance items such as ROC ,AUC , sensivity , specifity, percission of network by scripting? my goal is to automation all of process about network(creating, training , and reporting performance) thanks

Answer 1

[behruz]

Hi Greg thanks again for your response my goal is to generate all above mentioned numbers automatically by script coding and save them into a output data set automatically is it possible?

it seems all above function introduced in your reply is graphical.

Answer 2

[Greg Heath]

The operations help, doc and type can be used to find details of any function.

 If you type 

 net = patternnet  %NO SEMICOLON!

 you will find the plot functions

plotFcns: {'plotperform', plottrainstate, ploterrhist, plotconfusion, plotroc }

Most of what you want can be obtain from looking at the code of one or more of these functions. For example,

 help plotconfusion
 doc  plotconfusion
 type plotconfusion

reveals the additional functions confusion and roc

In addition, you can search both the NEWSGROUP and ANSWERS with these terms to view posts.

Hope this helps.

Thank you for formally accepting my answer

Greg

[Greg Heath]

I think all of that info is available from what I have written above.

[Greg Heath]

P.S. You accepted your own comment as the correct answer.

Please change