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

Table 8 Comparison of models by MSE over all noise disturbance indicators based on a relevant noise measurement separated by ’-’

Indicator	Nullmodel	XGB_xy	XGB_ogc	STAR	Corr	Pred	Time
Road_high-road	0.0665	0.0635	0.0635	0.0635	0.94	0.00	95.76
Rail_high-rail	0.0114	0.0109	0.0109	0.0109	0.91	0.00	96.62
Air_high-air	0.0357	0.0314	0.0313	0.0313	0.96	0.00	95.88
Road_medhigh-road	0.2234	0.2060	0.2056	0.2058	0.96	0.10	93.73
Rail_medhigh-rail	0.0780	0.0671	0.0667	0.0671	0.96	0.60	93.31
Air_medhigh-air	0.1593	0.1295	0.1286	0.1286	0.98	0.00	93.42
Road_high_gt50-road_pwrw	0.0328	0.0320	0.0320	0.0320	0.91	0.00	96.43
Road_medhigh_gt50-road_pwrw	0.1557	0.1465	0.1462	0.1467	0.95	0.34	94.11
Road_high_sm50-road_gw	0.0519	0.0501	0.0501	0.0502	0.92	0.20	96.73
Road_medhigh_sm50-road_gw	0.2045	0.1918	0.1914	0.1914	0.94	0.00	93.85
Road_high_gt50-road	0.0328	0.0320	0.0320	0.0320	0.90	0.00	95.94
Road_medhigh_gt50-road	0.1557	0.1472	0.1465	0.1467	0.93	0.14	92.50
Road_high_sm50-road	0.0519	0.0501	0.0501	0.0501	0.92	0.00	96.73
Road_medhigh_sm50-road	0.2045	0.1914	0.1911	0.1911	0.95	0.00	94.70

ISSN: 1476-072X