A Novel Feature Selection Scheme and a Diversified-Input SVM-Based Classifier for Sensor Fault Classification

Algorithm 1

Proposed feature selection algorithm.

INPUT: , Training dataset; , Number of features; , Number of classes; , Number of features to select by algorithm.

OUTPUT: , Matrix of vectors of features selected for each class .

1: for in do

2:  

3:  for in do

4:    values of feature from class

5:    values of feature from class

6:    mean of from class

7:    mean of from class

8:   Obtain and using Eq(8)

9:   Obtain , using Eq(4)

10:   Obtain and using Eq(4)

11:   Obtain and using Eq(3)

12:   Obtain using Eq(2)

13:  end for

14:  Indices of maximum values of

15: end for