Skip to main content

Table 3 Measuring the accuracy of ten prediction models using nine summary statistics criteria

From: Time series analysis of malaria cases to assess the impact of various interventions over the last three decades and forecasting malaria in India towards the 2030 elimination goals

Models (n = 10)

Summary statistics (n = 9)

ME

RMSE

MAE

MPE

MAPE

MASE

ACF1

AIC

BIC

Linear

Training

− 2.71e−11

305,647.6

224,512.4

− 6.74

17.31

1.34

0.65

  

Test

− 4.33e + 05

436,429.1

433,277.9

− 248.19

248.19

2.59

   

Quadratic

Training

2.72e− 11

250,728.0

189,867.5

− 1.92

11.90

1.13

0.66

  

Test

8.41e + 04

148,494.2

122,609.9

49.43

70.07

0.73

   

Cubic

Training

− 2.52e− 11

262,603.7

191,917.9

− 2.08

11.82

1.15

0.67

  

Test

1.38e + 05

197,841.6

155,947.6

80.50

90.17

0.93

   

Moving average

Training

− 70,280.17

183,319.75

145,459.7

− 10.08

13.57

1.05

0.51

  

Test

5797.66

11,713.42

10,721.0

2.98

6.02

0.08

   

Loess

Training

− 7111.25

47,878.78

21,823.28

− 0.65

1.34

0.32

− 0.12

  

Test

342,636.37

355,731.07

342,636.37

197.43

197.43

5.06

   

Simple Exponential

Training

− 67,907

216,932

164,871

− 10.29

15.6

1.05

0.25

  

Test

− 2,009,565

2,015,929

209,565

− 1147.84

1147.84

  

932.35

936.84

Double Exponential

Training

− 68,756

216,914

164,812

− 10.56

15.36

0.97

0.23

  

Test

− 2,113,922

2,122,632

2,113,922

− 1213.69

1213.69

  

932.34

936.83

ARIMA(1,2,2)

Training

− 47,088

197,391

157,307

− 3.96

11.62

0.94

− 0.06

  

Test

− 43,021

64,631

57,451

− 24.39

32.57

0.34

 

774.15

779.48

Holt's additive

Training

− 54,609

213,525

162,359

− 5.72

11.74

0.97

0.24

  

Test

24,779

77,470

63,311

14.82

35.48

0.38

 

852.06

857.67

Holt's multiplicative

Training

− 42,161

225,610

172,941

− 6.12

13.15

1.03

0.35

  

Test

− 86,303

90,667

86,303

− 49.43

49.43

0.51

 

855.31

860.92

  1. ME: mean error; RMSE: root mean square error; MAE: mean absolute error; MPE: mean percent error; MAPE: mean absolute percent error; MASE: mean absolute scaled error; ACF: auto correlation function; AIC: akaike information criterion; BIC: bayesian information criterion
Back to article page

Malaria Journal

ISSN: 1475-2875

Contact us