Random walk

updated: 2022-11-16
#network-science/random-walks

Random walks are instrumental to network analysis. It provides theoretical foundation of community detection, centrality analysis, routing, and embedding.

Random walks in a simplest form

A simplest form is random walks in undirected and unweighted networks. An agent called a walker at a node and moves to node randomly chosen from the neighbors of node . This process is a markov process of the first order, i.e., the walker determines the next node based on the current node and has no memory of nodes previously visited.

Variants

Random walks with teleportation

Non-backtracking walks

Degree-avoiding

Self-avoiding walks

Node2Vec walks

Fairness-aware walks

Review