Use of machine learning techniques for identifying ischemic stroke instead of the rule-based methods: a nationwide population-based study

Table 1 Demographic characteristics of the study cohorts

Characteristic	Development cohort		Validation cohort		Interval validation cohort
Characteristic	n	%	n	%	n	%
Patients(n, %)	19,773	100.00	4944	100.00	6180	100.00
Age(years, mean ± SD)	67.42 ± 20.19		67.58 ± 19.94		67.36 ± 20.02
Gender(male, %)	9156	46.31	2318	46.89	2879	46.59
Income(n, %), medical aid	465	2.35	104	2.10	155	2.51
1	2054	10.39	543	10.98	638	10.32
2	2349	11.88	590	11.93	729	11.80
3	3748	18.96	942	19.05	1200	19.42
4	4398	22.24	1077	21.78	1389	22.48
5	6635	33.56	1668	33.74	2042	33.04
Death(n, %)	3731	18.87	919	18.59	1142	18.48
Rule-based method*(n, %)	18,065	91.36	4534	91.71	5663	91.63
Ischemic stroke (n, %)	1996	10.09	513	10.38	636	10.29

^*The rule-based method is for individuals who are hospitalized with a diagnosis of I63 and have received anti-platelet therapy and anti-coagulant therapy within 30 days of diagnosis; SD standard deviation

ISSN: 2047-783X