Deep neural networks predict HOMA-IR from wearables plus blood biomarkers with R²=0.5 and auROC=0.80 on N=1,165 participants, performing better in obese/sedentary subgroups and reproducing on an independent N=72 cohort.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Insulin Resistance Prediction From Wearables and Routine Blood Biomarkers
Deep neural networks predict HOMA-IR from wearables plus blood biomarkers with R²=0.5 and auROC=0.80 on N=1,165 participants, performing better in obese/sedentary subgroups and reproducing on an independent N=72 cohort.