how to confusion matrix convert to heatmap style?

how to confusion matrix convert to heatmap style?

I want to process the CNN-based modulation classification example to evaluate the learning performance in a new environment.

The confusionchart(a,b) function currently in use does not use axis and colorbar, and the color of the placed box is also not customizable.

In this regard, I would appreciate if you could solve it.

Answer 1

[dpb]

m=[923     4    21     8     4     1     5     5    23     6
     5   972     2     0     0     0     0     1     5    15
    26     2   892    30    13     8    17     5     4     3
    12     4    32   826    24    48    30    12     5     7
     5     1    28    24   898    13    14    14     2     1
     7     2    28   111    18   801    13    17     0     3
     5     0    16    27     3     4   943     1     1     0
     9     1    14    13    22    17     3   915     2     4
    37    10     4     4     0     1     2     1   931    10
    20    39     3     3     0     0     2     1     9   923];
>>  classLabels=categorical(...
     {'airplane',     'automobile' ,     'bird',     'cat',     'deer' ,
     'dog',     'frog' ,     'horse' ,     'ship',     'truck'});
heatmap(classLabels,classLabels,m)

produces

You can experiment about what is modfiable and see if joy ensues...

[youngwoo Oh]

Amazing!!! thx dpb!!