8000 Cross validation and tools for undersampling and oversampling · Issue #1362 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]
Skip to content
Cross validation and tools for undersampling and oversampling #1362
New issue
New issue
Closed
Closed
Cross validation and tools for undersampling and oversampling#1362
Labels
New Featuremodule:model_selection
@amueller

Description

@amueller
amueller
opened on Nov 13, 2012

In learning with unbalanced classes, it is often good to stratify the data (not as in StratifiedKFold) by making the classes the same size, either by oversampling or undersampling.
I think it would be good to have both a utility function that does that once, and a cross-validation object that does that differently for all folds.

Metadata

Metadata

Assignees

No one assigned

    Labels

    New Featuremodule:model_selection

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions
      0