DAISY LabSharif University of Technology
Publications

Joint Inference of Diffusion and Structure in Partially Observed Social Networks Using Coupled Matrix Factorization

Maryam Ramezani, Aryan Ahadinia, Amirmohammad Ziaei Bideh, Hamid R. Rabiee

ACM Transactions on Knowledge Discovery from Data, 18 July 2023

arXiv

Abstract

Access to complete data in large-scale networks is often infeasible, so missing data is an unavoidable issue in the analysis of real-world social networks. In this paper, a model is learned from partially observed data to infer unobserved diffusion and structure networks. To jointly discover omitted diffusion activities and hidden network structures, we develop a probabilistic generative model called DiffStru. The interrelations among links of nodes and cascade processes are utilised via learning coupled low-dimensional latent factors. Besides inferring unseen data, latent factors such as community detection may also aid in network classification problems. Experiments on simulated independent cascades over LFR networks and on real datasets including Twitter and Memetracker show that the proposed method successfully detects invisible social behaviours, predicts links, and identifies latent features.

In short

DiffStru treats the two things we usually cannot observe — who is connected to whom, and which diffusion events actually happened — as a single joint problem rather than two separate ones.

The model couples low-dimensional latent factors across the link structure and the cascade process, so evidence about one side constrains the other. The same factors turn out to be useful beyond imputation: they carry community structure that can be reused for classification.

Cite this paper

@article{ramezani2023joint,
  title = {{Joint Inference of Diffusion and Structure in Partially Observed Social Networks Using Coupled Matrix Factorization}},
  author = {Maryam Ramezani and Aryan Ahadinia and Amirmohammad Ziaei Bideh and Hamid R. Rabiee},
  journal = {ACM Transactions on Knowledge Discovery from Data},
  year = {2023},
  eprint = {2010.01400},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2010.01400}
}

Related work