Neural Network , bad results

Hi all,

I ran a simple decision tree on a credit dataset (with fake data), basically for predicting the outcome of a credi evaluation process, with the class variable beiong valued Bad/Good.

The model scores reasonably well.

I tried to did the same and train a neural network MLP Percepton on the same data set, and what I've got is a 100% error rate on the confusion matrix.

I included the workflow I've build , can someone have a quick look at it and tell me where is the error ? Thanks a lot in advance!!

PS: I'm a newbie with NN....

 

Giovanni  

 

Are you sure you are columns to be compared to generatethe confusion matrix are right? They certainly seem to be incorrect in the Scorer node at present.

 

See attached screenshot with Configure Dialog and Scorer confusion matrix. Looks alright to me.

Are you sure that the columns to be compared in Scorer to generate the confusion matrix are right? They certainly seem to be incorrect in the Scorer node at present. See ealier post.

Thanks a lot! That was exactly the problem!

Giovanni