Table 2 Best performing feature sets for “at admission” and “post admission” models
From: Machine learning-based mortality prediction models for smoker COVID-19 patients
Feature set | Method | Number of features | Features | |
|---|---|---|---|---|
At admission | 7 | Feature Importance using Gradient Boosting | 20 | Age, Oxygen Saturation Percent, Chronic Kidney Disease, Respiratory Rate, Diastolic Blood Pressure, Systolic Blood Pressure, BMI, Average Daily Used Cigarettes, Pantoprazole, Cancers, Hypertension, Abnormal Lung Signs, Drug History, Sex, Total Ling Involvement Percent, Hospitalization in a 14-day period prior to admission, Current Smoking, Cardiovascular Disease, Chronic Obstructive Pulmonary Disease, Diabetes |
Post admission | 8 | Physician Opinion | 24 | Age, BMI, Systolic Blood Pressure, Diastolic Blood Pressure, Respiratory Rate, Oxygen Saturation Percent, Total Lung Involvement Percent, Sex, Current Smoking, History of Hookah consumption, Drug History, Fever, Dyspnea, Chest Pain, Diabetes, Hypertension, Cancers, Cardiovascular Disease, Chronic Kidney Disease, Chronic Obstructive Pulmonary Disease, Immunosuppressant Drugs, Duration of Intubation, Duration of Non-invasive ventilation, Admission in intensive care unit |