Skip to main content

Table 2 SVM evaluation with balanced accuracy, MCC, and F1 score

From: Classifying natural products from plants, fungi or bacteria using the COCONUT database and machine learning

 

Balanced acc.

MCC

F1

MAP4 SVM a,b

0.919 ± 0.005

0.879 ± 0.005

0.929 ± 0.003

ECFP4 SVM a,b

0.890 ± 0.005

0.827 ± 0.006

0.897 ± 0.003

RDKit AP SVM a,b

0.735 ± 0.005

0.592 ± 0.006

0.752 ± 0.004

Properties SVM a,c

0.758 ± 0.005

0.613 ± 0.007

0.761 ± 0.004

  1. aMean value and standard deviation (σ) of the five different test/training sets split of the fivefold cross-validation
  2. b1024 dimensions
  3. c11 properties: MW, Fsp3, HBD) and HBA, calculated logP with the Crippen method (AlogP), number of carbons, oxygen, and nitrogen, the total number of atoms, number of bonds, and topological polar surface area (TPSA)
Back to article page

Journal of Cheminformatics

ISSN: 1758-2946

Contact us