Skip to main content
. 2022 Jul 22;122:802–810. doi: 10.1016/j.ijid.2022.07.034

Figure 2.

Figure 2

(A) A LSTM network that takes multiple time points of clinical laboratory values and vital signs as input. (B) A network that takes as input the demographics and comorbidities values. (C) An LSTM network that combines both longitudinal feature values and the demographics and comorbidities values to predict hospital-acquired AKI onset in COVID-19 patients. AKI, acute kidney injury; LSTM, long-short-term memory.