Table 5 Chosen features of the classifiers used in the study

Multilayer Perceptron

Neural networks with backpropagation used for tuning the weights of a neural net based on the error rate (i.e. loss).

BNN

AutoBuild: true;

Learning rate: 0.3;

Momentum: 0.1;

Training time: 500

Hidden layers = 25

LibSVM

This library enables users to deal with One-class SVM, Regressing SVM, and nu-SVM. Many useful statistics are allowed including confusion matrix, precision estimation, ROC score.

LIBSVM

SVM Type: nu-SVC;

Kernel Type: radial basis function;

Nu: 0.g;

gamma: 0.1;

degree: 3

Normalize: true;

Probability Estimates: true

Logistic

Used for building and using a multinomial logistic regression model with a ridge estimator.

LOG

Debug: false;

MaxIts: −1;

Ridge: 1.0E-6

Random Forests

This classifier enables to create forest of random trees. It induces each constituent decision tree from a bootstrap sample of the training data

RF

Debug: false;

MaxDepth: 0;

Num of Features: 0;

Num of Trees: 10;

Seed: 1