A Novel Hierarchical Extreme Machine-Learning-Based Approach for Linear Attenuation Coefficient Forecasting

. 2023 Jan 30;25(2):253. doi: 10.3390/e25020253

Algorithm 2 ELM-Autoencoder Overview.

1:
Input: The training dataset matrix X: $(x_{i}) | x_{i} \in$ $R^{n}$ , i = 1, …, N; The number of hidden layer for sparse autoencoder: m;
2:
Output: Output weight vector of each layer: $β$ $_{i}$ ;
3:
Randomly generate hidden node matrix for original ELM $V_{(m + 1)}$ ,
4:
$O_{0} \overset{}{\leftarrow} X;$
5:
for (i = 0; i < x; i++) do
6:
Calculate hidden weight vector $β$ $_{i}$ by Algorithm 1 using $O_{i - 1}$ and $V_{i}$ as its parameters
7:
$O_{i}$ ← $O_{i - 1} \times β$ $_{i}$
8:
end for
9:
Compute output weight $β$ m + 1 by Equation (8) using $O_{m}$ and $V_{m}$ as its parameters;
10:
Return $β_{i}, i = 1, \dots, (m + 1)$ .