A study criticized for data leakage used anti-hypertensive-drug use as a feature to predict hypertension, even though that information would not be available when predicting the outcome for a new patient.
Sources
- Leakage and the reproducibility crisis in machine-learning-based science (cell.com)
- https://pmc.ncbi.nlm.nih.gov/articles/PMC7880048/ (pmc.ncbi.nlm.nih.gov)
- https://www.jmir.org/2021/2/e10969 (jmir.org)
- https://arxiv.org/pdf/2412.03491 (arxiv.org)
- https://www.jmir.org/2018/1/e22/ (jmir.org)
More in Science
- According to Italy's Passi surveillance, about 6% of adults aged 18-69 report depressive symptoms.
- In the Di Francesco et al. (2024) study of female mice, the lifespan-extension effect was largest with 40% chronic caloric restriction, followed in order by 20% chronic caloric restriction, 2-day intermittent fasting, and 1-day intermittent fasting.
- The JAMA Neurology study found that sleeping nine hours or more per night was linked to cognitive problems, particularly impaired decision-making.
- The effects of age-distribution shifts on unemployment in the U.S. are larger than a mechanical shift-share model predicts, indicating amplifying indirect effects.
- BICEP and the Keck Array are a series of cosmic microwave background (CMB) experiments deployed near the South Pole.
- The transition from non-life to life has never been observed experimentally.