0
$\begingroup$

I am training a SVM model for binary classification. For this, I have split the train and test datasets in an 80:20 ratio. Then I standardized the training and test data separately and tuned the hyperparameters of the SVM model using grid-search.

Using grid-search, I have found the following best parameters:

{'kernel': 'rbf', 'gamma': 0.0001, 'C': 100}

Now, I have trained the SVM model with these hyperparameters using 80% training data. After that, I have tested the trained SVM model using 20% test data and got acc_score, f1, precision, recall, auroc etc.

My supervisor told me to include the mean and standard deviation of accuracy, F1-score, precision, recall, and AUROC metrics. I have passed all the test data at once to the trained SVM model.

I don't know how to compute the standard deviation of accuracy, F1-score, precision, recall, and AUROC metrics. Any hints or suggestions will be most welcome.

$\endgroup$

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.