How can i put the conditional statement ( If-Then-else) on the specific elements (Rows/columns) of a matrix in Matlab Using Neural Network Tool?
3 views (last 30 days)
Show older comments
Safdar Hayat Khaksar
on 15 Oct 2015
Edited: Safdar Hayat Khaksar
on 30 Oct 2015
How can i put the conditional statement ( If-Then-else) on the specific elements (Rows/columns) of a matrix in Matlab Using Neural Network Tool? Basically I am working on classification using neural Network tool in matlab. My out put only shows in matrix like : output =
1.2723
0.0123
0.0022
But I want the output in statement like "Class1" Or "Class2"
i.e. If the matrix is like this then output is this....
For example: I have two different matrices like: output1 =
1.2723
0.0123
0.0022
output2 =
0.2460
-0.0521
1.1062
%I want to set the ranges of the elements on behalf of some conditions like:
If The first element of the first row of matrix Output1 is >= 0.6 and <=1.5
AND
If The 2nd element of the 2nd row of matrix Output1 is >= -0.49 and <=0.49
AND
If The 3rd element of the 3rd row of matrix Output1 is >= -0.49 and <=0.49
THEN
Display = Class1
%for Matrix Output2:
I want to set the ranges of the elements on behalf of some conditions like:
If The 1st element of the 1st row of matrix Output2 is >= -0.49 and <=0.49
AND
If The 2nd element of the 2nd row of matrix Output2 is >= -0.49 and <=0.49
AND
If The 3rd element of the 3rd row of matrix Output2 is >= 0.6 and <=1.5
THEN
Display = Class2
Please anyone can send me the code/explanation. how can I set these condition using If-Then-else or any other statement?
Regards:
Safdar Hayat
+923335260136
Skype:
jugnu2008.safdar
Iqra University Islamabad, Pakistan
0 Comments
Accepted Answer
Greg Heath
on 17 Oct 2015
If you have 4 classes, your target matrix columns should be columns of eye(4). The relationship between class indices 1:4 and the target matrix is
trueclassindices = vec2ind(target)
target = ind2vec(trueclassindices)
Example:
>> trueclassindices = [ 1 1 2 4 3 4 ]
>> target = full( ind2vec(trueclassindices))
target = 1 1 0 0 0 0
0 0 1 0 0 0
0 0 0 0 1 0
0 0 0 1 0 1
If
output =[ 0.5142 0.49114 0.18198 0 0.25288 0.13838
0.03653 0.1455 0.4168 0.14434 0.18459 0.066341
0.31642 0.22294 0.26658 0.38809 0.54807 0.35987
0.13285 0.14041 0.13464 0.46757 0.014469 0.43541]
then
assignedclasses = 1 1 2 4 3 4
Hope this helps.
Thank you for formally accepting my answer
PS. The unit sum output is a consistent estimate of the input-conditional posterior class probabilities (See any pattern-recognition text).
Greg
3 Comments
Greg Heath
on 18 Oct 2015
I understand your question.
My answer is to change your target to have {0,1} unit vectors, modify your code and train a new classifier.
Hope this helps.
Greg
More Answers (0)
See Also
Categories
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!