Broader Perception for Local Community Identification

The local identification of communities in a network can be much more effective than partitioning the entire network when only a small portion of a large network is of interest. A local approach is also favored when it is difficult to obtain information about the entire network, the world wide web is a prime example. Such local identification algorithms typically evaluate nodes outside the local community that are directly connected to the local community as potential community members without the need for knowledge about the entire network. But while being a recognized and relevant technique in community identification, the quality of local community identification lags behind their global counterparts that use the entire network. In order to improve the evaluation of community candidates in local algorithms we propose the use of contextual information beyond direct neighbors. This will decrease the gap between relevant network knowledge of global and local methods while remaining a local approach. Benchmarks on synthetic networks show our approach increases the quality of locally identified communities in general and a decrease of the dependency on specific source nodes.