leilamr / wine-mlp

Wine Classification with Neural Network

Wine Classification with Neural Network

The problem

Classify wines from three wineries by thirteen attributes.

Network settings:

Train = 70%, Validation = 15% and Testing = 15%
Number hidden of nodes = 4 
Epochs = 1000
Trainng Function = trainlm
Transfer Function (layer 1) = logsig
Trasnfer Function (layer 2) = purelin 

Accuracy = 99.4%

Confusion Matrix