Table 3 The top 20 features selected by the different approaches with bold text indicates commonly selected features.

From: A comparison of machine learning algorithms and traditional regression-based statistical modeling for predicting hypertension incidence in a Canadian population

Random survival forest relative importance

Statistical equivalent signature

Harrel’s C-index/Somers’ Dxy rank correlation

Lasso Cox coefficients/variable importance feature

Univariate Cox p-values

Top 20 features

Systolic blood pressure

Systolic blood pressure

Systolic blood pressure

Waist-hip ratio

Systolic blood pressure

Diastolic blood pressure

Body mass index

Diastolic blood pressure

Cardiovascular disease

Age

Body mass index

Cardiovascular disease

Body mass index

Diabetes

Body mass index

Waist-hip ratio

Age

Age

Job schedule

Waist-hip ratio

Diabetes

Diabetes

Waist-hip ratio

Residence

Diastolic blood pressure

Cardiovascular disease

Job schedule

Job schedule

Total household income

Total household income

Age

Total household income

Working status

Alcohol Consumption

Job schedule

Job schedule

Working status

Highest education level completed

Sex

Working status

Working status

Total sleep time

Total household income

Depression

Cardiovascular disease

Total household income

Ethnicity

Residence

Highest education level completed

Diabetes

Residence

Waist-hip ratio

Total sleep time

Body mass index

Highest education level completed

Total sleep time

Highest education level completed

Smoking status

Family history of hypertension

Residence

Highest education level completed

Residence

Sex

Systolic blood pressure

Sex

Family history of hypertension

Smoking status

Diabetes

Age

Smoking status

Physical activity, quartiles

Marital status

Total physical activity time

Diastolic blood pressure

Total sleep time

Smoking status

Alcohol consumption

Cardiovascular disease

Total physical activity time

Family history of hypertension

Total physical activity time

Family history of hypertension

Alcohol consumption

Total sitting time

Total physical activity time

Depression

Sex

Family history of hypertension

Ethnicity

Vegetable and fruit consumption

Ethnicity

Physical activity, quartiles

Vegetable and fruit consumption

Working status

Alcohol consumption

Sex

Diastolic blood pressure

Marital status

Vegetable and fruit consumption

Marital status