Fig 7. Global view of feature impact on mortality prediction.

Features are ranked in descending order of their accountability for the prediction. Each dot in the visualization represents one datapoint of a feature. Its color is related to the real data value: high value in red and low value in blue. BMI: body mass index. Cancer: history of cancer. dBP: first measurement of diastolic blood pressure at ER. DM: history of diabetes mellitus. Heart: history of heart disease. HTN: history of hypertension. ICU_adm: whether a patient was admitted into ICU or not. LOS: length of stay in hospital. Pulse: first measurement of pulse at ER. Pressors: if a patient received vasopressor treatment. Respiration: first measurement of respiration rate at ER. sBP: first measurement of systolic blood pressure at ER. Steroid: whether a patient received steroid treatment. Steroid_dur: duration of steroid treatment. tAC: if a patient received anticoagulation treatment. tAC_dur: duration of anti-coagulation treatment. Temp: first measurement of temperature at ER. Vented: where a patient was mechanically vented or not.