Predicting Potential Risk: Cerebral Stroke via Regret Minimization

Table 2

Comparison of method accuracy across different feature scales.

Feature number	Method name
	Chi-square test (%)	F test (%)	Mutual information (%)	RMFS (%)
1	(acPayStyle, 70.83)	(dfHypertension, 69.73)	(dfHypertension, 69.73)	(age_4, 69.73)
2	(acPayStyle, acJob, 69.73)	(dfHypertension, dfSportsLack, 69.85)	(dfHypertension, acPayStyle, 71.75)	(age_4, dfGlycuresis, 71.38)
3	(acPayStyle, acJob, age_4, 72.18)	(dfHypertension, dfSportsLack, acPayStyle, 71.26)	(dfHypertension, acPayStyle, acJob, 71.45)	(age_4, dfGlycuresis, acPayStyle, 72.92)
5	(acPayStyle, acJob, age_4, age_6, glu_high, 73.59)	(dfHypertension, dfSportsLack, acPayStyle, acJob, ldlc_low, 70.10)	(dfHypertension, acPayStyle, acJob, dfSportsLack, lsDrink, 70.10)	(age_4, dfGlycuresis, acPayStyle, age_5, ldlc_low, 74.69)
10	73.90	74.51	74.63	76.16
20	75.00	76.04	74.14	76.35
30	75.86	75.74	76.35	77.27
40	76.16	76.35	75.55	77.02
50	75.98	76.41	76.59	76.72
60	76.96	76.96	76.35	77.08

The bold values indicate the minimum values in each column. The smaller the value is, the better the value is.