Non-rule-based Classifiers   Top 100 Genes Top 150 Genes Top 200 Genes Top 250 Genes Naïve Bayes 67% 61% 58% 58% Logistics 54% 64% 64% 67% Multi-Layer Perceptron 64% 64% 61% 64% RBF Network 64% 61% 67% 58% Simple Logistic 67% 70% 70% 70% SMO 61% 67% 67% 64% Random Tree 71% 71% 70% 69% Rule-based Classifiers   Top 100 Genes Top 150 Genes Top 200 Genes Top 250 Genes Decision Table 73% 74% 70% 74% JRIP 74% 68% 77% 74% PART 73% 73% 73% 73% NNGE 68% 67% 70% 70% Data-adaptive partition classifiers   Top 100 Genes Top 150 Genes Top 200 Genes Top 250 Genes Gradient-based partitioning 74% 74% 74% 74%

Table 3: Pre-processed Dataset (After Feature Selection, 31 samples, and 2 classes (Normal and Disease).