Tag: data leakage

A systematic review of 60 strawberry detection studies finds that 35 percent applied data augmentation before dataset splitting, a controlled ...

An audit of an AI oral health screening prototype found that benchmark duplication did not change aggregate scores, but label ...

A new dual-branch AI framework called TBFuseFormer detects tuberculosis on chest X-rays with 99.39 percent held-out test accuracy after removing ...

A dual-path attention fusion network combining Swin Transformer and frozen DINOv2 features reached 99.24% accuracy on a leakage-screened hold-out test ...

A leakage-controlled nested cross-validation study found handwriting-based machine learning can classify Alzheimer's disease with high descriptive accuracy, but task selection ...

A sweeping methodological review published in Frontiers in Oncology argues that the explosion of public-omics research in gastric cancer has ...

A four-color offset-printing study found stable optimal-density estimates across validation designs despite major changes in model performance and rankings.

Researchers built a synthetic, Japan-calibrated medical claims benchmark with traceable fraud counterfactuals and a frozen single-access evaluation, reporting strong anti-fraud ...

A 240-run reevaluation of potato disease classifiers shows that DINOv2 visual clusters poorly match published leaf identities and that grouped ...

A new leakage-controlled machine learning framework achieves strong sentiment classification on social media data while revealing that engagement prediction collapses ...