Skip to main content

Table 1 Performance comparison of classifiers

From: G4Boost: a machine learning-based tool for quadruplex identification and stability prediction

 

Accuracy

F1-score

Precision

Recall

AUROC

Average

SD

Average

SD

Average

SD

Average

SD

Average

SD

XGBoost

0.938

0.002

0.964

0.001

0.959

0.006

0.969

0.004

0.976

0.002

NeuralNet

0.934

0.002

0.962

0.001

0.964

0.009

0.959

0.011

0.974

0.002

RandomForest

0.933

0.003

0.961

0.002

0.956

0.006

0.966

0.003

0.963

0.005

KNN

0.928

0.003

0.958

0.002

0.955

0.007

0.961

0.005

0.937

0.010

CART

0.928

0.005

0.958

0.003

0.958

0.006

0.958

0.005

0.919

0.016

LR

0.927

0.003

0.958

0.002

0.949

0.007

0.968

0.005

0.966

0.003

LDA

0.907

0.005

0.947

0.003

0.932

0.004

0.962

0.004

0.941

0.005

NBayes

0.790

0.019

0.862

0.014

0.990

0.005

0.764

0.022

0.943

0.015

  1. Five-fold cross-validation performance metrics of eight classifiers for the prediction of G4 structure folding probability. Average is devoted to average score and std is to standard deviation for the five-fold cross-validation sets. AUROC is the area under the receiver operating characteristic curve