Community Discovery Algorithm Based on Topological Information and Feature Attributes
A novel weighted algorithm aimed at improving the accuracy of community detection algorithms is proposed.The algorithm integrates network topology information and node attribute information,utilizing spec-tral clustering techniques to reveal potential community structures.It is capable of adaptively learning the contribution of different attributes to the detection task and accordingly assigning weights,while employing soft threshold operators and L1 regularization terms to reduce the adverse impact of"noise"attributes on de-tection precision.The effectiveness of the algorithm in community detection applications has been validated through the tests on numerical simulations and real-world datasets.
community detectionspectral clusteringstochastic block modelweight update
NETL
NSTL
万方数据
