A novel model and algorithm for discovering the potential Web communities based on tensor decomposition

Based on the analysis of the typical models and algorithms for the discovering the latent Web communities, we give a novel model and its algorithm using tensor CP decomposition for discovering the latent Web communities. The model considers three signatures of the Web pages, which adds a text score of the Web page except for the authority score and hub score. We also present an iteration algorithm for tensor decomposition. Experiment tests show that our model and algorithm are effective.