Skip to main content

Table 3 Cross-validation results of the four models

From: Comparison of new computational methods for spatial modelling of malaria

Fold

Model

RMSE

Correlation

% points with absolute error less than

0.05

0.1

0.2

10-fold

INLA

0.181

0.235

69.211

76.316

86.053

GPBoost

0.127

0.646

52.632

73.947

93.158

SpRF

0.132

0.641

69.474

80.263

91.053

FRK

0.125

0.661

57.105

76.579

92.105

50-fold

INLA

0.124

0.683

69.474

80.789

87.895

GPBoost

0.11

0.751

65.789

80.789

90

SpRF

0.121

0.702

67.632

78.947

90.526

FRK

0.123

0.702

66.053

79.211

90.263

  1. Boldface denotes the best score in each column for each of the cross-validation experiments