Fig. 1

RWRF and RWRNF overview. a We constructed a multiplex network in which corresponding samples of \(S_{1}\) and \(S_{2}\) are connected. An imaginary particle starts a random walk from seed node A. The number next to the node indicates the probability of walking to the node. After several iterations, the stationary probability distribution will be obtained. RWRF utilizes the stationary probability distribution representing similarities between A and other nodes in the multiplex network. Here, we set α = 0.5. b Unlike RWRF, RWRNF connects \(S_{1}\) and \(S_{2}\) differently. For the sake of clarity and brevity, we only draw the connections from A in \(S_{1}\) to other nodes in \(S_{2}\). Here, we set m = 2, α = 0.5, β = 0.5. Note that edges with similarity less than 0.5 are omitted