Skip to main content

Table 4 Comparison of the accuracy of different classifiers using 2 known biomarker genes and our selection of 6 genes on Ma et al. and Loi et al. data

From: A voting approach to identify a small number of highly predictive genes using multiple classifiers

Classifier Ma et al.data Loi et al.data
  2 genes 6 genes 2 genes 6 genes
C4.5 60.00% 100% 75.64% 80.77%
C4.5 with boosting (ADABoost) 70.00% 100% 66.67% 82.05%
C4.5 with bagging 70.00% 100% 67.95% 75.64%
Naïve Bayes 60.00% 100% 74.36% 74.36%
Naïve Bayes with boosting 60.00% 80.00% 74.36% 77.95%
Naïve Bayes with bagging 60.00% 100% 75.64% 75.64%
LMT 70.00% 100% 76.92% 79.49%
NBTree 80.00% 80.00% 75.64% 82.05%
Random Forest 60.00% 100% 74.36% 75.38%
Random Forest with boosting 70.00% 100% 67.95% 74.36%
Random Forest with bagging 70.00% 100% 74.36% 71.79%
k-NN 70.00% 100% 73.08% 71.79%
Logistic Regression 70.00% 100% 76.92% 74.36%
ANN 60.00% 100% 74.36% 76.67%
SVM 60.00% 100% 74.36% 74.36%
\