Table 2 Average performance of various ML classifiers, along with standard deviation, across the 10 testing experiments for the Glider scenario, \(\mathscr{D}^g\) (columns 2 through 4), Satellite scenario \(\mathscr{D}^s\) (columns 5 through 7), and Glider + Satellite scenario, \(\mathscr{D}\) (columns 8 through 10)

From: Machine learning for modeling North Atlantic right whale presence to support offshore wind energy development in the U.S. Mid-Atlantic

 

Glider, \(\mathscr{D}^g\)

Satellite, \(\mathscr{D}^s\)

Glider + Satellite, \(\mathscr{D}\)

Accuracy

F1 Score

AUC

Accuracy

F1 Score

AUC

Accuracy

F1 Score

AUC

LR

0.961 ± 0.006

0.198 ± 0.068

0.625 ± 0.035

0.965 ± 0.013

0.000 ± 0.000

0.560 ± 0.036

0.972 ± 0.006

0.265 ± 0.093

0.624 ± 0.041

SVM

0.946 ± 0.004

0.216 ± 0.039

0.785 ± 0.025

0.885 ± 0.014

0.133 ± 0.020

0.733 ± 0.030

0.923 ± 0.009

0.231 ± 0.034

0.856 ± 0.027

kNN

0.950 ± 0.006

0.298 ± 0.037

0.736 ± 0.050

0.967 ± 0.005

0.431 ± 0.053

0.792 ± 0.032

0.961 ± 0.005

0.373 ± 0.039

0.791 ± 0.039

RF

0.975 ± 0.007

0.374 ± 0.071

0.778 ± 0.030

0.965 ± 0.005

0.411 ± 0.040

0.883 ± 0.017

0.975 ± 0.003

0.524 ± 0.040

0.886 ± 0.032

AdaBoost

0.972 ± 0.004

0.377 ± 0.052

0.815 ± 0.045

0.982 ± 0.003

0.615 ± 0.058

0.866 ± 0.028

0.987 ± 0.002

0.675 ± 0.054

0.904 ± 0.027

± GBoost

0.971 ± 0.005

0.359 ± 0.066

0.823 ± 0.030

0.983 ± 0.003

0.641 ± 0.048

0.888 ± 0.029

0.986 ± 0.003

0.649 ± 0.048

0.891 ± 0.028

MLP

0.969 ± 0.011

0.334 ± 0.091

0.779 ± 0.036

0.964 ± 0.006

0.358 ± 0.057

0.803 ± 0.032

0.971 ± 0.008

0.421 ± 0.080

0.863 ± 0.024

CNN

0.977 ± 0.009

0.344 ± 0.091

0.764 ± 0.036

0.943 ± 0.015

0.267 ± 0.064

0.801 ± 0.033

0.955 ± 0.010

0.345 ± 0.072

0.842 ± 0.025

ResNet

0.943 ± 0.002

0.223 ± 0.108

0.769 ± 0.043

0.933 ± 0.038

0.186 ± 0.065

0.735 ± 0.022

0.935 ± 0.031

0.289 ± 0.096

0.862 ± 0.031

  1. Bold-faced values denote best performance for each metric under each scenario.
Back to article page