Utilizing imbalanced electronic health records to predict acute kidney injury by ensemble learning and time series model

摘要

Acute Kidney Injury (AKI), a sudden loss of kidney functions, is a shared complication in the Intensive Care Unit (ICU) patients [1]. The incidence of AKI usually causes a significant drain on medical resources and increases patients’ morbidity and mortality [2]. It is noteworthy that timely detection and management can effectively reverse patients’ conditions. Therefore, the early prediction of AKI helps physicians give patients timely medical interventions and is critical for improving patients’ outcomes.