Hanczar09a
Supplementary results:
Comparison of classification algorithms for microarray data
Blaise Hanczar (1), Edward Dougherty (2,3)
(1) CRIP5, University Paris Descartes, 45 rue des Saint-Peres, 75006 Paris, FRANCE
(2) Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843, USA
(3) Computational Biology Division, Translational Genomics Research Institute, Phoenix, AZ 85004, USA
line 1: True error of the two algorithms
line 2: Comparison of estimated and true difference of error
line 3: Décompositon of the variane into external and internal variance
line 4: Comparison of the true and internal variance of the estimators
| Linear Gaussian model, 100 examples, 500 uncorrelated features, 20 features selection by t-test score | ||||
| experiment ID | 1 | 2 | 3 | 4 | |
| 3NN vs SVM | 3NN vs LDA | QDA vs CART | 1NN vc SVM | |
| 1 | True error rates | ||||
| 2 | Comparison of estimated and | ||||
| 3 | Decomposition of the variance | variance decomposition | variance decomposition | variance decomposition | variance decomposition |
| 4 | Comparison true / estimated variance | variance comparison | variance comparison | variance comparison | variance comparison |
| Nonlinear Gaussian model, 100 examples, 500 uncorrelated features, 20 features selection by t-test score | ||||
| experiment ID | 5 | 6 | 7 | 8 | |
| 3NN vs SVM | 3NN vs LDA | QDA vs CART | 1NN vc SVM | |
| 1 | True error rates | ||||
| 2 | Comparison of estimated and | ||||
| 3 | Decomposition of the variance | variance decomposition | variance decomposition | variance decomposition | variance decomposition |
| 4 | Comparison true / estimated variance | variance comparison | variance comparison | variance comparison | variance comparison |
| Nonlinear Gaussian model, 100 examples, 500 correlated features, 20 features selection by t-test score | |||
| experiment ID | 9 | 10 | 11 | |
| 3NN vs SVM | 3NN vs LDA | QDA vs CART | |
| 1 | True error rates | |||
| 2 | Comparison of estimated and | |||
| 3 | Decomposition of the variance | variance decomposition | variance decomposition | variance decomposition |
| 4 | Comparison true / estimated variance | variance comparison | variance comparison | variance comparison |
| Nonlinear Gaussian model, 500 correlated features, 20 features selection by t-test score, 3NN vs LDA | |||
| experiment ID | 9 | 10 | 11 | |
| 200 examples | 300 examples | 500 examples | |
| 1 | True error rates | |||
| 2 | Comparison of estimated and | |||
| 3 | Decomposition of the variance | variance decomposition | variance decomposition | variance decomposition |
| 4 | Comparison true / estimated variance | variance comparison | variance comparison | variance comparison |
References of the articles published in Bioinformatics and BMC Bioinformatics on new methods of classification for microarray data: here
