Predicting Hidden Links and Missing Nodes in Scale-Free Networks with Artificial Neural Networks

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autor principal: Pran, Rakib Hassan
Formato: Preprint
Publicado: 2021
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917353727459328
author Pran, Rakib Hassan
author_facet Pran, Rakib Hassan
contents There are many networks in real life which exist as form of Scale-free networks such as World Wide Web, protein-protein interaction network, semantic networks, airline networks, interbank payment networks, etc. If we want to analyze these networks, it is really necessary to understand the properties of scale-free networks. By using the properties of scale free networks, we can identify any type of anomalies in those networks. In this research, we proposed a methodology in a form of an algorithm to predict hidden links and missing nodes in scale-free networks where we combined a generator of random networks as a source of train data, on one hand, with artificial neural networks for supervised classification, on the other, we aimed at training the neural networks to discriminate between different subtypes of scale-free networks and predicted the missing nodes and hidden links among (present and missing) nodes in a given scale-free network. We chose Bela Bollobas's directed scale-free random graph generation algorithm as a generator of random networks to generate a large set of scale-free network's data.
format Preprint
id arxiv_https___arxiv_org_abs_2109_12331
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Predicting Hidden Links and Missing Nodes in Scale-Free Networks with Artificial Neural Networks
Pran, Rakib Hassan
Social and Information Networks
Machine Learning
There are many networks in real life which exist as form of Scale-free networks such as World Wide Web, protein-protein interaction network, semantic networks, airline networks, interbank payment networks, etc. If we want to analyze these networks, it is really necessary to understand the properties of scale-free networks. By using the properties of scale free networks, we can identify any type of anomalies in those networks. In this research, we proposed a methodology in a form of an algorithm to predict hidden links and missing nodes in scale-free networks where we combined a generator of random networks as a source of train data, on one hand, with artificial neural networks for supervised classification, on the other, we aimed at training the neural networks to discriminate between different subtypes of scale-free networks and predicted the missing nodes and hidden links among (present and missing) nodes in a given scale-free network. We chose Bela Bollobas's directed scale-free random graph generation algorithm as a generator of random networks to generate a large set of scale-free network's data.
title Predicting Hidden Links and Missing Nodes in Scale-Free Networks with Artificial Neural Networks
topic Social and Information Networks
Machine Learning
url https://arxiv.org/abs/2109.12331