Partial Classifier Chains with Feature Selection by Exploiting Label Correlation in Multi-Label Classification

. 2020 Oct 10;22(10):1143. doi: 10.3390/e22101143

Algorithm 1. The training phase of the classifier chain (CC) algorithm

Input:

X_{t r a i n}

Output:

C_{j}, j = 1, 2, \dots, Q

Steps

for $j \in 1, 2, \dots, Q$ ;
do /*the jth binary transformation and training*/;
$X_{t r a i n}^{'} \to {}$ ;
for $(x_{i}, y_{i}) \in X_{t r a i n}$ , $i = 1, 2, \dots, m$ ;
do $X_{t r a i n}^{'} \leftarrow X_{t r a i n}^{'} U {((x_{i}, l_{1}, \dots, l_{j - 1}), l_{j})}$ ;
$C_{j} : X_{t r a i n}^{'} \to l_{j} \in {0, 1}$ /*train $C_{j}$ to predict binary relevance of $l_{j}$ */.