Premium
Incorporating domain ontology information into clustering in heterogeneous networks
Author(s) -
Huang Yue
Publication year - 2021
Publication title -
wiley interdisciplinary reviews: data mining and knowledge discovery
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.506
H-Index - 47
eISSN - 1942-4795
pISSN - 1942-4787
DOI - 10.1002/widm.1413
Subject(s) - cluster analysis , computer science , ontology , data mining , domain (mathematical analysis) , knowledge extraction , focus (optics) , object (grammar) , similarity (geometry) , information retrieval , artificial intelligence , image (mathematics) , mathematics , mathematical analysis , philosophy , epistemology , physics , optics
Clustering of structure‐rich heterogeneous information networks composed of multiple types of objects and relationships, which has become a challenge in data mining. Most of the existing clustering heterogeneous network methods focus on the internal information of the dataset while ignoring the domain knowledge outside the dataset. However, in real‐world scenarios, domain knowledge can often offer valuable information for clustering. In this study, we propose a three‐layer model OntoHeteClus, which is able to cluster multitype objects in star‐structured heterogeneous networks by considering both the dataset itself and the background information quantified via the ontology. OntoHeteClus first evaluates the similarity between central objects according to formalized domain ontology information, based on which central objects are subsequently clustered. Finally, attribute objects are clustered according to the central object clustering result. A numerical example is presented to illustrate the modeling concept and working principle of the proposed method, and experiments on a real‐world dataset demonstrate the effectiveness of the proposed algorithms. This article is categorized under: Technologies > Structure Discovery and Clustering Algorithmic Development > Structure Discovery