Development of decision tree classification algorithms in predicting mortality of COVID-19 patients

Table 4 Ranking from low to high importance of each attribute in predicting COVID-19 mortality using decision tree indices

Attributes	Information gain	Gini index	Gain ratio	Average ranks
Chest pain	4	2	2	2.67
Olfactory	3	1	4	2.67
Abdominal pain	5.5	3.5	5.5	4.83
Asthma	5.5	3.5	5.5	4.83
Nausea	9	7	3	6.33
Anorexia	10	8	1	6.33
Seizure	7	5	11	7.67
Liver disease	8	6	13	9
Vomiting	11	9	10	10
Muscular pain	12	12	7	10.33
Headache	13	11	9	11
Cough	15	14	8	12.33
Dizziness	17	16	15	16
Body temperature	1.5	34	16	17.17
Immunodeficiency	14	15	24	17.67
Drug	19	17	18	18
Cigarettes	20	19	19	19.33
Diarrhea	22	20	17	19.67
HIV	18	10	35	21
Gender	29	23	12	21.33
Hemiparesis	16	18	30	21.33
ALT	1.5	40	23	21.5
Fever	30	27	14	23.67
Pregnancy	21	13	37	23.67
Cancer	27	25	28	26.67
HTN	31	29	21	27
Respiratory distress	33	30	20	27.67
Loss of taste	26	21	39	28.67
Hematologic disease	23	22	42	29
Heart disease	32	31	26	29.67
Hemiplegia	24	24	44	30.67
ESR	34	38	22	31.33
Type 2 diabetes	36	32	27	31.67
Nervous disorder	28	26	41	31.67
Dermatitis	25	28	48	33.67
Hemoglobin	37	41	25	34.33
AST	41	42	29	37.33
RBC	39	43	31	37.67
Kidney disease	35	33	47	38.33
Length of hospitalization	42	36	40	39.33
Loss of consciousness	38	35	46	39.67
BUN	48	35	38	40.33
ALP	40	48	34	40.67
CRP	46	44	33	41
WBC	45	46	32	41
NLR	43	47	36	42
O2 sat	44	37	45	42
Age	47	39	43	43
ICU hospitalization	49	49	49	49
Intubation	50	50	50	50

ISSN: 1865-1380