Recognition of Multiple Imbalanced Cancer Types Based on DNA Microarray Data Using Ensemble Classifiers

Pseudocode 2

Pseudocode description of the ensemble learning algorithms based on feature subspace and counter voting rule for classifying imbalanced multiclass cancer microarray data.

Input: Training set 𝑇; Feature set 𝐹; Size of feature space 𝐾; Size of

Feature subspace 𝐷; Number of classes 𝐶; Number of feature subspace

𝐿; Baseline learner 𝐼; One test sample 𝑥^′

Output:  ℎ(𝑥^′) which is the class label of the test sample 𝑥^′

Process:

(1)  for 𝑖=1 : 𝐶

(2)  {

(3)  Label the samples of 𝑖th class as positive and the rest samples as

   Negative;

(4)  External 𝐿 diverse training subsets by feature subspaces generation

   algorithm (see Pseudocode 1);

(5) for 𝑗=1 : 𝐿

(6)  {

(7)   Train imbalanced base classifier 𝐼_𝑖,𝑗 by training subset 𝑇_𝑖,𝑗 using

   THR or RUS, abbreviated as EnSVM-OAA(THR) and

    EnSVM-OAA(RUS), respectively.

(8)  }

(9)  }

(10) for 𝑖=1 : 𝐶

(11) {

(12)  for 𝑗=1 : 𝐿

(13)  {

(14)    Use 𝐼_𝑖,𝑗 to classify the test sample 𝑥^′;

(15)  }

(16) Calculate the value of Counter_𝑖;

(17) }

(18) output ℎ(𝑥^′) by (10)