Data and Code for: Machine-learning estimation of vancomycin AUC24 with temporal validation and practical sampling-window optimization.

This repository contains the public analysis materials used to support the development and evaluation of a single-sample machine-learning workflow for estimating vancomycin AUC24. The deposit includes the analysis code, data structure documentation, and public-ready templates needed to reproduce the main analytical steps reported in the manuscript. The workflow is designed for a single-sample therapeutic drug monitoring setting and includes data cleaning, cohort construction, feature engineering, model development, model comparison, final-model evaluation, manuscript-oriented figure generation, and sampling-window analysis. To protect patient privacy and institutional confidentiality, raw individual-level clinical data are not deposited in this record. Public materials are limited to de-identified or template-based resources that allow readers to understand the data structure, analytical workflow, variable definitions, and computational steps used in the study. Example or template input files may be included solely to illustrate formatting and execution requirements. This record is intended to support transparency, reproducibility of the analytical workflow, and reuse of the modeling framework in similar single-sample vancomycin AUC studies.