In this work, we consider the community detection problem on hypergraph networks. It often occurs that network information comes with additional attributes on nodes, which could be used to improve our understanding of the network structure. We thus propose a probabilistic generative model that is able to use the information about higher-order interactions as well as the node attributes to infer the structure of the network. We demonstrate a variety of cases where using our model provides a significant advantage compared to the methods that do not use any attribute information or the methods that infer network structure from attributes alone. The proposed method is able to identify automatically if the attributes are informative and discard them otherwise. We show the benefits of using our model on the link prediction task when the given attribute is informative. The model comes with an efficient implementation that allows it to generalize to hypergraphs of large size both in terms of the number of nodes and number of edges.

In this work, we consider the community detection problem on hypergraph networks. It often occurs that network information comes with additional attributes on nodes, which could be used to improve our understanding of the network structure. We thus propose a probabilistic generative model that is able to use the information about higher-order interactions as well as the node attributes to infer the structure of the network. We demonstrate a variety of cases where using our model provides a significant advantage compared to the methods that do not use any attribute information or the methods that infer network structure from attributes alone. The proposed method is able to identify automatically if the attributes are informative and discard them otherwise. We show the benefits of using our model on the link prediction task when the given attribute is informative. The model comes with an efficient implementation that allows it to generalize to hypergraphs of large size both in terms of the number of nodes and number of edges.

Inference and Community Detection in Hypergraphs: Incorporating Node Attributes

BADALYAN, ANNA
2023/2024

Abstract

In this work, we consider the community detection problem on hypergraph networks. It often occurs that network information comes with additional attributes on nodes, which could be used to improve our understanding of the network structure. We thus propose a probabilistic generative model that is able to use the information about higher-order interactions as well as the node attributes to infer the structure of the network. We demonstrate a variety of cases where using our model provides a significant advantage compared to the methods that do not use any attribute information or the methods that infer network structure from attributes alone. The proposed method is able to identify automatically if the attributes are informative and discard them otherwise. We show the benefits of using our model on the link prediction task when the given attribute is informative. The model comes with an efficient implementation that allows it to generalize to hypergraphs of large size both in terms of the number of nodes and number of edges.
2023
Inference and Community Detection in Hypergraphs: Incorporating Node Attributes
In this work, we consider the community detection problem on hypergraph networks. It often occurs that network information comes with additional attributes on nodes, which could be used to improve our understanding of the network structure. We thus propose a probabilistic generative model that is able to use the information about higher-order interactions as well as the node attributes to infer the structure of the network. We demonstrate a variety of cases where using our model provides a significant advantage compared to the methods that do not use any attribute information or the methods that infer network structure from attributes alone. The proposed method is able to identify automatically if the attributes are informative and discard them otherwise. We show the benefits of using our model on the link prediction task when the given attribute is informative. The model comes with an efficient implementation that allows it to generalize to hypergraphs of large size both in terms of the number of nodes and number of edges.
community detection
probabilistic models
hypergraph analysis
File in questo prodotto:
File Dimensione Formato  
dissertation.pdf

accesso aperto

Dimensione 4.95 MB
Formato Adobe PDF
4.95 MB Adobe PDF Visualizza/Apri

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/62023