MATLAB Answers

0

how can i train my network for test and validation data?

Asked by Ahsen Feyza Dogan on 16 Jan 2019
Latest activity Answered by Greg Heath
on 16 Jan 2019
x is the my input, y is the desired output then I split x as validation and test data y as well. How should be my algorithm? I want to see how weights are changing for each epoch.
x=[5,1,1,1,2,1,3,1,1;5,4,4,5,7,10,3,2,1;3,1,1,1,2,2,3,1,1;6,8,8,1,3,4,3,7,1;4,1,1,3,2,1,3,1,1;8,10,10,8,7,10,9,7,1;1,1,1,1,2,10,3,1,1;2,1,2,1,2,1,3,1,1;2,1,1,1,2,1,1,1,5;4,2,1,1,2,1,2,1,1;1,1,1,1,1,1,3,1,1;2,1,1,1,2,1,2,1,1;5,3,3,3,2,3,4,4,1;1,1,1,1,2,3,3,1,1;8,7,5,10,7,9,5,5,4;7,4,6,4,6,1,4,3,1;4,1,1,1,2,1,2,1,1;4,1,1,1,2,1,3,1,1;10,7,7,6,4,10,4,1,2;6,1,1,1,2,1,3,1,1;7,3,2,10,5,10,5,4,4;10,5,5,3,6,7,7,10,1;3,1,1,1,2,1,2,1,1;5,10,10,8,5,5,7,10,1;1,1,1,1,2,1,3,1,1;5,2,3,4,2,7,3,6,1;3,2,1,1,1,1,2,1,1;5,1,1,1,2,1,2,1,1;2,1,1,1,2,1,2,1,1;1,1,3,1,2,1,1,1,1];
y=[2;2;2;2;2;4;2;2;2;2;2;2;4;4;4;2;2;4;2;4;4;2;4;2;4;2;2;2;2]
x_test=x(1:20,:)
x_validation=x(21:end,:)
y_test=y(1:20,:)
y_validation=y(21:end,:)

  0 Comments

Sign in to comment.

2 Answers

Answer by Greg Heath
on 16 Jan 2019

Where is your training data?
The MATLAB default is a random choice of trn/val/tst = 0.7/0.15/0.15
I choose candidate values of H to avoid overfitting and, 10-to-20 random initial sets of weights.
There are zillions of my examples in BOTH COMP.SOFT-SYS.MATLAB & ANSWERS
Hope this helps.
Thank you for formally accepting my answer
Greg

  1 Comment

Actually I am new at this field so I confused that I am on the right way. I am trying to do split my data set x into test and validation then using single and multilayer perceptron get output as 2 or 4 depends on my input. Could you briefly explain what am I suppose to do?
Thanks

Sign in to comment.


Answer by Greg Heath
on 16 Jan 2019

The MATLAB default AUTOMATICALLY splits the data into TRAINING + VALIDATION+TEST SUBSETS in the ratios
0.7/0.15/0.15.
For regression and curvefitting see the documentation and examples using the commands
help fitnet
and
doc fitnet
For classification and pattern recognition use
help patternnet
and
doc patternnet
For examples just add "greg" to the commands.
Hope this helps
Greg

  0 Comments

Sign in to comment.