Antibiotic Resistance Feedback 100k

This upload contains the full technical documentation and a free evaluation sample (the first 1,000 patients of every table, both scenarios: 33,774 rows per scenario, 67,548 in total) — not the full dataset. Full dataset (200,000 synthetic patients, 18 Parquet tables, 6.89 million rows): https://sentineldata.com.ua/dataset/antibiotic-resistance-feedback-100k Sample contents (SAMPLE.7z, verified by unpacking and counting): Table (in scenario_A/ and scenario_B/) Rows × Cols per scenario Content cohort 1,000 × 16 Admission covariates, infection focus, reported pathogen and susceptibility, first dose observed 8,000 × 8 Eight creatinine draws per patient with the dose chosen after each draw observed_outcomes 1,000 × 11 KDIGO stage, dose reductions, drug given truth 1,000 × 67 Parameters, true course, cause of each creatinine rise, errors of the dosing estimate truth_traj 7,774 × 15 Hourly true states of the 46 sampled patients that belong to the 5,000-patient trajectory subset (169 rows each) truth_counterfactual 2,000 × 25 Same patients under two fixed alternative dosing rules truth_noloop 1,000 × 22 Same patients with the renal feedback layer switched off truth_exposure_daily 7,000 × 19 True daily exposure, burden and renal state regime_arms 5,000 × 27 Five dosing regimes on the same patients with the same noise sample passport PDF, ABOUT_PARQUET.txt — Sample data passport; short ... Full description with all tables is on the Wiki page of this project. Full dataset and purchase: https://sentineldata.com.ua/dataset/antibiotic-resistance-feedback-100k . Zenodo copy: https://doi.org/10.5281/zenodo.23435912