How to calculate accuracy, F1 score & entropy?

Here is my data ""

Now I have to split this dataset into 70% training set & 30% test set....

Then I have to calculate accuracy, F1 score & entropy using some classifiers. They are Decision tree, knn, svm

How can I do this? Please help

[Yazan]

This is not a question, but rather an assignment. See Mathworks examples on the Statistics and Machine Learning Toolbox.

Answer 1

[Ram Patro]

The data you have provided does not contain class label information. When you have the class label vector 'classLabel', you can partition data using cvpartition function.

per = 10; % Training percentage 
cv = cvpartition(classLabel,HoldOut=1-(per/100)); 

'cv.training' lists all the training location indices that you can use to partition the data. Similarly '~cv.training' lists all the testing location indices.

For classification, you can refer to the examples:

• fitctree function for decision tree classifier.
• fitcknn function for K- neareset neighbour classifier
• fitcsvm function for binary models of SVM classification
• fitcecoc function for multiclass models of SVM classification.

After obtaining your classification results, you can refer:

• confusionmat and confusionchart for preparing the confusion matrix.
• crossentropy function to calculate cross-entropy loss
• this for other binary classification measures.