Input: $\left\{E^1,\cdots ,E^M\right\}$ .

Output: The edge weight matrix of the target layer.

Calculate the weights of each layer:

For $m\in \left\{b,c,d\right\}$ .

The ratio of overlapping edges between each layer and the initial state of the target layer is taken as the weight of each layer.

$w_m=likelihood\left(Linkin{L}^a|Linkin{L}^m\right)$

Calculate the edge weights of the target layer:

For $e\in E^a$ , #Math_23#, indicates e is the edge from i to j.

$w_e=rate+{\displaystyle \sum _{m\in \left[b,c,d\right]}{w}_m\times linkexist\left(e\right)}$ $linkexist\left(e\right)=\{\begin{array}{l}1,e\text{existsinlayer}m,\\ 0,e\text{doesnotexistinlayer}m\text{.}\end{array}$

The rate part indicates the impact of layer (a), the formula is

$rate={\log }_{100}\left(s_{out}\left(i\right)\right)+{\log }_{100}\left(s_{in}\left(j\right)\right)$