Table 1 Model fit metrics based on k-fold cross-validation using observed datasets on Papua New Guinea

From: Estimating small area population from health intervention campaign surveys and partially observed settlement data

Metric

In-Sample Cross-Validation

Out-of-Sample Cross-Validation

BHM (Count)

TSBHM (Count)

BHM (Density)

TSBHM (Density)

BHM (Density)

TSBHM (Density)

MAE

3.191

2.251

0.062

0.005

0.670

0.007

RMSE

5.734

4.213

1.891

0.220

14.716

0.172

Abias

3.142

2.123

0.054

0.003

0.665

0.003

CC

0.991

0.991

0.992

0.792

0.866

0.862

  1. The predictive abilities of the proposed TSBHM and the conventional BHM methods were evaluated using a k-fold cross-validation technique. Model performances were evaluated and compared based on Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Absolute bias (Abias), and the Pearson correlation coefficient (CC). Lower values of MAE, RMSE, Abias and higher values of CC indicated higher predictive ability.
Back to article page