Skip to main content

Table 2 Summary of features used for prediction in HeMo and WoON

From: A machine learning approach to small area estimation: predicting the health, housing and well-being of the population of Netherlands

Feature

Categories/median (min.–max.)

% missing

Age

50 (18–108)

0.0

Sex

Male

0.0

Female

Ethnicity

Netherlands

0.0

Morocco

Turkey

Suriname

Netherlands Antilles

Other non-western

Other western

Marital status

Single

0.6

Married

Divorced

Widowed

Education (descriptions Table 13)

Basis

37.8

VMBObk

VMBOgt

MBO23

MBO4

HAVO-VWO

HBO-WO-BAC

HBO-WO-M/PhD

Household type

Single person household

0.0

Unmarried without children

Married without children

Unmarried with children

Married with children

Single parent family

Other

Household size

2 (1–10)

0.0

Household income source

Wage

2.4

Wage director/shareholder

Self-employed

Unemployment benefit

Social assistance benefit

Disability benefit

Old-age pension

Other benefit

Student loan

Property income

Home ownership

Homeowner

2.0

Rental no allowance

Rental with allowance

Household income (percentile)

63 (1–100)

2.4

Household assets (percentile)

56 (1–100)

2.4

Neighborhood address density

71 (1–100)

0.0

x-coordinate (m)

140,348 (13,666–277,711)

0.0

y-coordinate (m)

453,926 (306,922–611,538)

0.0