Skip to main content

Table 6 Comparison of the classifier performance using i) a variable subset of our 7 genes, ii) a set of 17 genes identified by ROC with Markov Blanket [9], iii) a set of 17 genes identified by LAD [15], iv) a set of 70 genes identified by van 't Veer [3] and v) a set of 231 genes identified by van 't Veer [3]

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

Classifier Subsets of our 7 genes Our 7 genes Set of 17 genes[9] Set of 17 genes[15] Set of 70 genes[3] Set of 231 genes[3]
C4.5 with boosting 91.67% 84.52% 68.42% 59.52% 54.76% 76.19%
C4.5 84.52% 84.52% 68.42% 57.90%* 42.11%* 73.68%*
k-NN 80.36% 77.38% 74.21% 63.16%* 63.16%* 78.94%*
Logistic Regression 81.55% 77.38% 73.68% 73.68%* 47.37%* 73.68%*
ANN 77.38% 76.19% 84.21% 84.21%* 42.11%* 73.68%*
SVM 83.33% 76.19% 79.47% 63.16%* 57.90%* 73.68%*
  1. *Results adopted from Alexe et al. [15].
\