Skip to main content
. 2021 Apr 16;23(4):e24120. doi: 10.2196/24120

Figure 1.

Figure 1

Architectural overview of the developed model using the recurrent neural network (RNN). (A) The sliding window approach for handling sequential data and (B) the 2-stage hierarchical model comprising models 1 and 2. Model 1 could predict the occurrence of acute kidney injury (AKI) in the next 7 days from the present, and model 2 could provide the predicted value of serum creatinine (Cr) for the next 3 days.