inferCommunitiesDOLP: Community detection, "dropout" label propagation

In sverchkov/CommunityInference: Community Inference for Weighted Graphs

Description

An implementation of community detection by label propagation in an undirected weighted graph based on Raghavan, Albert, Kumara. Phys Rev E 76, 036106 (2007)

Usage

inferCommunitiesDOLP(unique_edges, async_prop = 0.5, check_unique = F)

Arguments

unique_edges	a data frame with columns a, b, weight representing the connections between nodes. We assume undirected graph, and therefore b < a.
async_prop	proportion of nodes to update before synchronous update
check_unique	whether to check edges data frame for uniqueness

Details

This version has "dropout" behavior – asynchronysity is implemented by only updating a random subset of nodes

Value

a data frame with two columns, node–node id (taken from a, b of input) and label–unique cluster/community ID.

Author(s)

Yuriy Sverchkov

sverchkov/CommunityInference documentation built on May 7, 2019, 7:42 a.m.