Skip to main content

Table 1 Overview of MaxEnt models

From: Do Marmorkrebs, Procambarus fallax f. virginalis, threaten freshwater Japanese ecosystems?

Trained on

Sample size

beta

Threshold

Parameters

AICc score

Training AUC

Test AUC (30%)

Procambarus fallax

37

1.0

MTSPS = 0.284

12

874.2029424

0.9773

 

Procambarus fallax

37

1.5

MTP = 0.284

12

903.8337866

0.9764

0.9766 (0.0043)

Marmorkrebs populations

23

1.0

MTSPS = 0.033

22

not applicable

0.9844

 

Marmorkrebs populations

23

1.5

MTP = 0.014

18

760.5960693

0.9813

0.9823 (0.0146)

Marmorkrebs populations and Sapporo individual

24

1.0

MTSPS = 0.045

26

not applicable

0.9869

 

Marmorkrebs populations and Sapporo individual

24

1.5

MTP = 0.030

23

not applicable

0.9836

0.9723 (0.0212)

P. fallax and Marmorkrebs populations

58

1.0

MTSPS = 0.328

39

1829.92309

0.9673

 

P. fallax and Marmorkrebs populations

58

1.5

MTP = 0.027

27

1660.981878

0.9623

0.9595 (0.0119)

P. fallax, Marmorkrebs populations, and Sapporo individual

59

1.0

MTSPS = 0.321

37

1882.800434

0.9640

 

P. fallax, Marmorkrebs populations, and Sapporo individual

59

1.5

MTP = 0.020

24

1722.570932

0.9586

0.9537 (0.0148)

  1. Training AUC is the AUC obtained in the 1-replicate run used for model selection and MESS generation. Test AUC is the average AUC (and standard deviation) obtained in 10 runs that used a 30% of the presence points to calculate this statistic; only models considered superior based on reduced complexity without training AUC reduction were projected and test AUC calculated. MTSPS – minimum training sensibility plus specificity; MTP – minimum training presence; not applicable – AICc cannot be calculated if the model have more parameters than occurrence points [29].