Community Detection in Social Networks via Laplacian Eigenmaps

Abstract

Community detection in the domain of complex social networks is a fundamental task that reveals the underlying structure and organization of the network. The majority of network data give properties that describe nodes and show a specific structural relationship between them.By making use of the node features that are available, hidden communities within an observed network can be found.This paper presents a method for detecting communities using Laplacian Eigenmaps, a spectral dimensionality reduction technique that preserves local geometric structures. We provide a comprehensive mathematical formulation of Laplacian Eigenmaps and demonstrate its effectiveness in identifying communities through a series of experiments on real-world social network datasets. The results highlight the ability of the proposed Laplacian Eigenmaps method to capture intrinsic graph properties, facilitating accurate community detection and offering valuable insights into network dynamics.