Fig. 2: Framing of the machine learning algorithm. | npj Digital Medicine

Fig. 2: Framing of the machine learning algorithm.

From: The Framing of machine learning risk prediction models illustrated by evaluation of sepsis in general wards

Fig. 2: Framing of the machine learning algorithm.

Eight different structures to framing the sepsis prediction model are shown. Fixed time to onset (a), sliding window (b), sliding window with dynamic inclusion (c), on clinical demand (d), sequential with prediction window (e), sequential with entire admission as prediction window (f), at event (g), and random time to onset (h). The top four structures (a–d) were explored, and the bottom four (e–h) were not. * indicates that the samples were collected from two different patients/admissions. EWS early warning score, SOFA sequential organ failure assessment scores, OR operating room.

Back to article page