Making a NN more accurate through an ROC curve

Hey, guys im testing the accuracy of my NN(MLP learner nodes) using an ROC curve node however Im getting 100% accuracy on the ROC how can I resolve this issue?

Thus how can i improve the accuracy of the NN?
Testing.knwf (48.0 KB)

Do you have data you can upload as well? It seems likely that your ROC Curve is misconfigured somehow, but I can’t open the configuration dialog to check.

If your data is confidential, screenshots of the config dialogs of the ROC Curve and Scorer would help.

Also, how large is your dataset?

2 Likes

Hi there @pizafile,

please take a look at this topic regarding sharing workflow examples:

Br,
Ivan

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.