Skip to main content

Table 5 Comparison of performance of models in terms of “early warning success” using the AUC metric

From: Predicting rice blast disease: machine learning versus process-based models

Model

Early warning Success – AUC metric

 

Kalochori 2016

Seville

2016

Kalochori

2015

Portugal

2015

Average of Normalized valuesb

Yoshino

28

46

38

N/Aa

0.74

WARM

38

N/Aa

18

12

0.77

M5Rules

38

88

15

12

0.80

LSTM NN

39

62

13

14

0.76

  1. aResults are not available for these locations, bValues are normalized columnwise then average taken for each row