Skip to content

Wrong crros validation #4

@tomer196

Description

@tomer196

Hi,
I think you did mistake in your cross-validation. If I understand correctly you just train on all the data and not separate the folds to different model.
I got low test AUC (0.69) and high AUC (0.98) on what you called 'cross_iter10' which is essentially the training AUC (Again, I hope I understand correctly the code). Is this the AUC you reported in the paper?

Tomer

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions