Multi-Population Genetic Algorithm for Multilabel Feature Selection Based on Label Complementary Communication

. 2020 Aug 10;22(8):876. doi: 10.3390/e22080876

Algorithm 1 Label Complementary multi-population genetic algorithm for multilabel feature selection

1:
Input: $D, m$ ; ▹ the multilabel dataset D, the number of sub-populations m
2:
Output: S; ▹ the final feature subset S
3:
$t \leftarrow 0$ ;
4:
$[P_{1} (t), \dots, P_{m} (t)] \leftarrow$ initialization $(m)$ ▹ use Algorithm 2
5:
for each sub-population $P_{k}$ do
6:
$v_{k} (t) \leftarrow$ evaluate $P_{k} (t)$ using D; ▹ compute fitness values via a fitness function
7:
$A_{k} (t) \leftarrow$ compute the label-specific accuracy matrix for individuals of $P_{k} (t)$ ; ▹ reuse the fitness function
8:
end for
9:
while (not termination-condition) do
10:
for each sub-population $P_{k}$ do
11:
$i n d^{c} \leftarrow$ communication $(P_{k} (t), A)$ ; ▹ use Algorithm 3
12:
$P_{k} (t + 1) \leftarrow$ update $(P_{k} (t), i n d^{c})$ ; ▹ use Algorithm 4
13:
$v_{k} (t + 1) \leftarrow$ evaluate $P_{k} (t + 1)$ using D;
14:
$A_{k} (t + 1) \leftarrow$ compute the label-specific accuracy matrix for individuals of $P_{k} (t + 1)$ ;
15:
$t \leftarrow t + 1$ ;
16:
end for
17:
end while
18:
$S \leftarrow$ the best feature subset so far;