Hello Fatemah,
When using machine learning technique like K-Cross Validation, the best practice is to:
- Start by splitting the data into training and test sets.
- Divide the training data into K folds.
- In each iteration, use K-1 folds to train the model and the remaining fold to validate it.
- After this process, a model with the best parameters is obtained, which can then be tested using the test set.
This approach ensures the model is evaluated on unseen data, providing a better assessment of its performance.
Alternate approach is to use the entire dataset as training data:
- Split the entire dataset into K folds, without a separate test set.
- Again, use K-1 folds for training and the remaining fold for validation.
- Here, since there is no separate test set, the validation fold in each iteration serves as the test fold, and the validation accuracy is used to evaluate the model.
Refer to below links for more information on:
- K-Cross Validation: https://www.analyticsvidhya.com/blog/2022/02/k-fold-cross-validation-technique-and-its-essentials/
- Implementation of K-Cross Validation in MATLAB: https://www.mathworks.com/support/search.html/answers/2149929-how-to-divide-image-datastore-into-training-set-validation-set-and-test-set-for-training-a-cnn-netw.html
