PU-Boost treats potentially undiagnosed hypertension as a positive-unlabeled learning problem. Elastic Net first performs global P/U stratification, a Random Forest then refines the uncertain region, and persistently ambiguous samples are rejected before final supervised learning on a reconstructed balanced set.
Open figure
Open figure
Open figure
Open figure
Open figure
Open figure
A single-center retrospective cohort study of enteric-release budesonide in Chinese patients with IgA nephropathy, using longitudinal mixed models to quantify proteinuria change while tracking kidney-function trajectories.
A concept-stage transformer framework for cfDNA methylation representation learning, extending read-level sequence context toward sample-level aggregation and prediction.
An early-stage data infrastructure project for through-silicon vias that turns heterogeneous literature, experimental and simulation evidence into standardized, provenance-aware multimodal records. Geometry, materials, process and operating conditions are linked to imaging, electrical, thermal, mechanical and reliability observations for benchmarking, surrogate modeling and simulation validation.
Click to enlarge
Click to enlarge
Click to enlarge
Data Science · Statistics
Computational Research
The University of Arizona
Department of Mathematics
Tucson, Arizona