notebook.community

Edit and run

Test Accuracy Rate Summary

Paper1 Selection Paper3 Selection Paper6 PCA Selection Paper6 PLS Selection Paper9 Selection Paper29 Selection
Paper1 classifiier 0.912 0.941 0.971 0.971 0.958 0.706
Paper 3 NN 0.971 0.941 0.912 0.941 1 0.912
Paper 3 SVM Linear 0.971 0.971 0.941 0.971 1 0.765
Paper 3 SVM Quadratic 0.971 0.882 0.971 0.971 1 0.912
Paper 3 Adaboost 0.912 0.912 0.971 0.971 0.958 0.941
Paper 6 logit 0.706 0.971 0.971 0.971 1 0.853
Paper 6 qda 0.971 0.912 0.941 0.971 1 0.853
Paper 9 nn 0.971 0.912 0.853 0.971 0.958 0.971
Paper 9 decision tree 0.912 0.912 0.971 0.971 0.917 0.735
Paper 9 bagging 0.971 0.912 0.971 0.971 0.958 0.735
Paper 9 bagging with CPD 0.941 0.912 0.971 0.971 0.917 0.794
Paper 9 FLDA 0.912 0.912 0.971 0.971 0.958 0.794
Paper 9 DLDA 0.941 0.912 0.971 0.971 0.958 0.765
Paper 9 DQDA 0.912 0.912 0.971 0.971 0.958 0.735
Paper 29 Bayesian Network 0.735 0.882 0.971 0.971 1 0.647
  • Test data in paper 9 only has 24 samples.
  • Paper6 pca/pls data only have 3 components.
  • paper 1, paper 6 and paper 9 variable selection perform overall better than others.
  • Bayesian network performs badly compared to other classifiers.
  • paper 29 variable selection is bad.

Content source: CompareML/CompareML-Manuscript

Similar notebooks:

notebook.community | gallery | about