How to feed a feature matrix for each class in multiclass classification?

Hello

I am using 'emd' for feature extraction. In my case, I will create 5 IMFs and extract 2 features from each so it would be a 5x2 matrix for each class label. How do I feed that to a multiclass classifier?

Answer 1

[Prince Kumar]

Hi,

You can feed the input matrix like you do it binary classifier or any classifier.

Please refer the following code for better understanding:

load fisheriris
Mdl = fitctree(meas,species)

Mdl = 
  ClassificationTree
             ResponseName: 'Y'
    CategoricalPredictors: []
               ClassNames: {'setosa'  'versicolor'  'virginica'}
           ScoreTransform: 'none'
          NumObservations: 150


  Properties, Methods

After you load the fisheriris dataset, you get meas (150x4) as input matrix and species(150x1) as target array.

Now you simply call "fitctree" to train a multiclass decision tree.

You can now use "predict" function to do the classification on the trained classifier.

X = [4.85,3.4,1.3,0.22];
label = predict(Mdl,X)

label = 1×1 cell array
    {'setosa'}

Hope this helps!