Skip to main content

Table 4 Model prediction accuracies based on cross-validation of the training set. RF-10: Random forest with 10 random decision trees, RF-20: Random forest with 20 random decision trees, SVM: default support vector machine, SVM-N: SVM with normalized features, MLP: default Multilayer perceptron, MLP-C: Multilayer perceptron with complex nodal architecture (described in methods)

From: Massive metagenomic data analysis using abundance-based machine learning

Model

Accuracy

RF-10

87.9

RF-20

89.7

SVM

43.1

SVM-N

32.8

MLP

63.7

MLP-C

55.2