Skip to main content

Table 1 Performance of 9 classifiers before and after feature selection

From: Identification of properties important to protein aggregation using feature selection

  560 features 7 features 10 features
  Accuracy MCC Accuracy MCC Accuracy MCC
SVM-linear 0.759 0.518 0.771 0.542 0.788 0.576
RF 0.748 0.497 0.737 0.474 0.782 0.564
GBM 0.754 0.509 0.718 0.436 0.797 0.593
RPART 0.717 0.435 0.751 0.502 0.729 0.457
NNet 0.754 0.507 0.734 0.468 0.780 0.558
PLS 0.740 0.479 0.788 0.578 0.782 0.565
KNN 0.762 0.530 0.763 0.524 0.763 0.528
NB 0.731 0.465 0.743 0.488 0.790 0.581
Ada 0.754 0.509 0.740 0.479 0.779 0.558
  1. The 7 features are selected by SVM-RFE and the 10 features are selected by RF-IS. The 10 fold cross validation of each classifier is conducted on dataset AP1.