bio-machine-learning-model-validation

What problem does it solve?

Nested cross-validation and stratified splits for unbiased evaluation of omics classifiers, helping prevent data leakage and overfitting in biomarker discovery.

Core Features & Use Cases

• Nested cross-validation with outer and inner folds to obtain robust performance estimates while tuning hyperparameters on biomedical data.
• Stratified K-Fold and group-aware splits to handle class imbalance and sample dependencies, with guidance to keep preprocessing inside the CV loop to prevent leakage.
• Practical use cases include validating classifiers for biomarker discovery in omics datasets and optimizing hyperparameters for biomedical models.

Quick Start

Run nested cross-validation with 5 outer folds and 3 inner folds to obtain a robust estimate of model performance.