The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
Neural Network With Multiple Outputs for Prediction
Hi,
I want a neural network with two outputs for prediction. I usually use the "set role" operator to set a lable, so the nerual network has one output. Now, how to set to get two outputs?
Thank you!
Tagged:
0
Answers
@Jeffery1 when you say "two outputs" do you mean two classes? Like the output of the ANN will be either "yes or no" or "churn or not churn?" Then selecting the one attribute column as a label is the right way to do it if you have the two classes in the column.
Thanks for your reply!
We usually use the neural network for prediction. I didn't mean two classes.The output of the ANN will be neither "yes or no" nor "churn or not churn" . I want it to be two numerical digits like "1234 and 5678" , that is to say the attributes of input and output are both numerical. So how to set to get two neurons of output layer ?
(PS:Please look at the attached picture.)
Thank you!
Jeffery
@Jeffery1 when you say the output is two numerical digits like '1234' and '5678', is that what expected out is to be? Or can it be like '2346', or '45', or '45745745746'? Do you want the ANN to output a predicted numerical value? Because in that case, you will not have two output nodes. Two output nodes (like your attached image) indicates two classes like "yes/no" or even "0/1"
Yes,i want the ANN to output a predicted numerical value. Most of the time, i want several predicted numerical values. Now ,i can only get one ,how to get more?
Thank you!
Jeffery
Hi @Jeffery1,
you are talking about multi-label learning like the MEKA library provides? E.g. http://meka.sourceforge.net/
I think this is not yet in RM. You would need to build independend models as of now. That might change soon.
Best,
Martin
Dortmund, Germany