If possible in neural networks turn off automatic data normalization?

답변 (1개)

Or course. You can change any default that you want.
[ x,t] = simplefit_dataset;
Now, what do you see when you enter each of these commands, ONE AT A TIME, without an ending semicolon? Look CAREFULLY at EACH result.
net = fitnet
input = net.inputs{1}
processFcns = net.inputs{1}.processFcns
processFcns2 = net.inputs{1}.processFcns{2}
output = net.outputs{2}
etc
Hope this helps.
Thank you for formally accepting my answer
Greg

댓글 수: 3

I deleted your question that was inappropriately put into an Answer Box.
The answer is
net = fitnet;
net.inputs{1}.processFcns{2} = {'removeconstantrows'};
net.outputs{2}.processFcns{2} = {'removeconstantrows'};
net.inputs{1}.processFcns = {'removeconstantrows'}; this was corrected! thanks
Good!
I think I was trying to eliminate the 2nd component directly. However I couldn't find a way to do it.
I don't know how I came up with the answer I posted.

댓글을 달려면 로그인하십시오.

카테고리

도움말 센터File Exchange에서 Deep Learning Toolbox에 대해 자세히 알아보기

질문:

2015년 6월 9일

편집:

2017년 2월 3일

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by