Naive Bayes-Guided Bat Algorithm for Feature Selection

Algorithm 1

BANB Algorithm.

(1) Initialize parameters: 𝐴,𝐴_min,𝐴_max,𝑟,𝑓_min,

  𝑓_max,𝑃_max,𝐼_max,𝑉_max,𝑉_min,Φ,𝛿,𝛾,𝛼

(2) Generate a swarm with 𝑃_max bats

(3) Calculate cost function for all bats

(4) Find the current best bat (𝑥_∗)

(5) While stop condition not met Do

(6)  For  𝑖 = 1 to  𝑃_max  Do

(7)   Frequency  𝑓_𝑖=𝑓_min+(𝑓_max−𝑓_min)𝛽

(8)   Velocity  𝑣^𝑡_𝑖=𝑣_𝑖^𝑡−1+(𝑥^𝑡_𝑖−𝑥_𝑖)𝑓_𝑖

(9)   If  𝑉_𝑖>𝑉_max  Then

           𝑉_𝑖=𝑉_max

     End-If

(10)       If  𝑉_𝑖<𝑉_min  Then

          (𝑉_𝑖=𝑉_min)

    End-If

(11)     Locations     𝑥^𝑡_𝑖=𝑥_𝑖^𝑡−1+𝑣^𝑡_𝑖

(12)    If (Rand>𝑟_𝑖 ) Then

(13)   calculate 𝜀𝐴^𝑡

(14)        If  𝜀𝐴^𝑡>𝐴_max  Then

       (𝜀𝐴^𝑡=𝐴_max)

     End-If

(15)        If  (𝜀𝐴^𝑡<𝐴_min)  Then

       (𝜀𝐴^𝑡=𝐴_min)

     End-If

(16)   Generate a local solution around the best

     solution (𝑥_∗) [𝑥_𝑔𝑏=𝑥_old+𝜀𝐴^𝑡]

(17)  End-If

(18)       Calculate 𝜀𝐴^𝑡

(19)  If  (𝜀𝐴^𝑡>𝐴_max)  Then

          𝜀𝐴^𝑡=𝐴_max

     End-If

(20)    If  (𝜀𝐴^𝑡<𝐴_min)  Then

      (𝜀𝐴^𝑡=𝐴_min)

     End-If

(21)   Generate a new solution around the current

     Solution  𝑥^𝑡_𝑖   [𝑥_𝑙=𝑥_old+𝜀𝐴^𝑡]

(22)    If  𝑥_𝑙≥𝑥_𝑔𝑏

        𝑓𝑥=𝑥_𝑙

        Else

        𝑓𝑥=𝑥_𝑔𝑏

    End-If

(23)      If  Rand<𝐴_𝑖&𝑓𝑓𝑥<𝑓𝑥_∗

(24)        accept the new solution

(25)        Increase 𝑟_𝑖𝑟_𝑖^𝑡+1=𝑟⁰_𝑖[1−exp(−𝛾𝑡)]

(26)        Decrease 𝐴_𝑖  [𝐴_𝑖^𝑡+1=𝛼𝐴^𝑡_𝑖]

(27)     End-If

(28)   End-For

(29)   Find the current best solution (𝑥_∗)

(30)  End-While