Competitive Particle Swarm Optimization for Multi-Category Text Feature Selection

. 2019 Jun 18;21(6):602. doi: 10.3390/e21060602

Algorithm 3 Competition function.

1:
input: $C, F, E_{c}, E_{f}, m_{c}, m_{f}$ ; ▹ fitness values for each group $E_{c}, E_{f}$
2:
output: $C, F, m_{c}, m_{f}$ ; ▹ changed particles via competitions $C, F$
3:
$w_{c} \leftarrow 0$ ;
4:
$l_{c} \leftarrow 0$ ;
5:
for $k = 1$ to $min ([m_{c}, m_{f}])$ do ▹ set the number of competitions
6:
if $max (E_{c}) > max (E_{f})$ then
7:
$w_{c} \leftarrow w_{c} + 1$ ; ▹ add one whenever evolution-based particle wins
8:
else if $max (E_{c}) < max (E_{f})$ and $m_{c} \neq l_{c} + 1$ then
9:
$l_{c} \leftarrow l_{c} + 1$ ; ▹ add one whenever evolution-based particle loses
10:
$j \leftarrow {j | \forall x \in {[1, m_{f}] \cap Z ∖ j}, E_{f} (j) > E_{f} (x)}$ ;
11:
$E_{f} (j) \leftarrow - \infty$ ; ▹ exclude winning particle at next competition
12:
end if
13:
end for
14:
for $k = 1$ to $w_{c}$ do
15:
$j \leftarrow {y | \forall x \in {[1, m_{f}] \cap Z ∖ y}, E_{f} (y) < E_{f} (x)}$ ;
16:
delete $F_{j}$ ; ▹ delete the particle with low fitness value
17:
$C_{e n d + 1} \leftarrow$ a new particle; ▹ use uniform distribution
18:
end for
19:
for $k = 1$ to $l_{c}$ do
20:
$j \leftarrow {y | \forall x \in {[1, m_{c}] \cap Z ∖ y}, E_{c} (y) < E_{c} (x)}$ ;
21:
delete $C_{j}$ ;
22:
$F_{e n d + 1} \leftarrow$ a new particle; ▹ use score vector for feature filter
23:
end for
24:
$m_{c} \leftarrow m_{c} + w_{c} - l_{c}$ ;
25:
$m_{f} \leftarrow m_{f} + l_{c} - w_{c}$ ;