Is there any limitation on the data ratio that is not suitable for cross-validation?
2 views (last 30 days)
Show older comments
Muhammad Usman Gul on 27 Dec 2021
I need to answer the reviewers' cross-validation after submitting my article to the Q1 journal. The scenario is that I did work on a hypothesis-based model and performed on linear classifiers. However, I also used augmentation techniques due to the scarcity of the original datasets (ratio in percentage 87:13), including a modified-augmentation technique. Overall, my article has novelty and a contribution, but I did not use cross-validation and testing due to the imbalanced dataset. I have extracted the relevant features using the filter and wrapper methods and shown the results on a training dataset.
Can anyone suggest a way to convince the reviewer why I did not use any cross-validation with appropriate logic? Is there any reference in which we could not perform cross-validation on a highly imbalanced dataset, even if augmentation techniques are used?
yanqi liu on 27 Dec 2021
yes，sir，may be plot roc and confuse matrix to do some analysis
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!