A toy example for learning the importance of meta-paths. (A) There are two node types and two edge types. (B) The miRNA–disease heterogeneous network, where the numbers on yellow edges and blue edges represent the similarities between the different nodes and the association relationships, respectively. (C) For the meta-paths with length equal to 2, we display all the paths from node |$m_{2}$| to |$d_{1}$|⁠. For each path, we first multiply the weights of its two edges to measure the total weight of the corresponding path. Then, MGCNSS could construct the weight matrix |$M_{c}^{2}$| in terms of the meta-paths with length 2. (D) The integrated result from |$m_{2}$| to |$d_{1}$| is shown in the weight matrix |$M_{c}^{2}$|⁠.