An Innovative Excited-ACS-IDGWO Algorithm for Optimal Biomedical Data Feature Selection

Algorithm 3. Pseudocode for the EACSIDGWO (Binary Version).

Input: labelled biomedical dataset D, MaxIter, ACS and IDGWO parameters value, number of host bird nests (), number of dimensions (features) , Lower bound () and Upper bound ()

Output: Best Fitness, Best Search Agent

1 for each nest i (i =1, 2...n) do

2 for each dimension j(j =1,2,…,d) do

3  =random number drawn from []

4  end

5 Convert continuous values ofto binary using Eq. (31), (32) and (36)

6 Train a classifier to evaluate the accuracy of the equivalent binary vector ofand store the value in

7 end

8 [~, Index] = Sort ()

9 =

10 =

11 =

12 =

13 =

14 =

15 =

16 While ()

17  for each nest i (i =1, 2...n) do

18   Calculateandusing

    Eq. (18) and (19), respectively

19   Generate a new cuckoo nest

    using Eq. (21)

20 Convert continuous values ofto binary using Eq. (31), (32) and (36)

21 Train a classifier to evaluate the accuracy of the equivalent binary vector ofand store the value in

22 if() then

23 =

24 =

25 end

26 end

27 Repeat step 8 to 15

28 for each nest i (i =1, 2...n) do

29 Calculateandusing Eq. (18) and (23), respectively

30 for each dimension j(j =1,2,…,d) do

31 Calculate coefficients and as shown in Equation (24) and Equation (10), respectively

32 Compute vectors, and using Equations (25), (26) and (27), respectively.

33 end

34 Convert continuous values ofto binary using Eq. (31), (32) and (36)

35 Consecutively, train a classifier to evaluate the accuracies of the equivalent binary vectors ofand store the value in, respectively.

36 Determineusing equations (28) and (29), respectively

37 end

38 Repeat step 8 to 15

39 Abandon a fraction of worst nests and generate new ones according to Equation (6)

40 Keep best solutions(or those nests with quality solutions)

41 Repeat step 8 to 15

 end

42 Best Search Agent=

43 Best Fitness=

Algorithm 3. Pseudocode for the EACSIDGWO (Binary Version).