Urine Routine Tests-Based Machine Learning and Urinalysis to Predict Acute Kidney Injury: URT-Based Machine Learning and Urinalysis to Predict AKI

Acute Kidney Injury (AKI) is a heterogeneous disease characterized by an increase in serum creatinine (Scr) or a decrease in urine output. It is a common critical illness with high morbidity, mortality, and clinical adverse events. In this study, we investigated the links that exist between data factors as well as target category features in urine routine tests (URT). Ten different machine learning algorithms were used to construct a classification prediction model for acute kidney injury, which can determine whether a patient suffers from acute kidney injury based on the patient's current physiological indicators. The dataset had a total of 33279 data (age $49\pm 24$ years, 48% women, 8914 AKI events) and 7 categories of features. Ten machine learning-based classifiers were trained and tested in a supervised learning environment. The maximum performance metric model has an accuracy of 0.934, precision of 0.93, recall of 0.90, and F1 score of 0.91. The experiments were tested for multi-algorithm comparisons, providing decision-making support on which algorithms need to be selected when studying such problems in the field of AKI prediction.