CST-Net: community-guided structural-temporal convolutional networks for popularity prediction

The ability to predict the popularity of online contents has important implications in a wide range of areas. The challenge of this problem comes from the inequality of the popularity of content and the numerous complex factors. Existing works fall into three main paradigms: feature-driven approache...

Full description

Saved in:
Bibliographic Details
Main Authors: Xuxu Zheng, Peng Bao, Lin Qi, Chen Tian, Huawei Shen
Format: Article
Language:English
Published: PeerJ Inc. 2025-06-01
Series:PeerJ Computer Science
Subjects:
Online Access:https://peerj.com/articles/cs-2931.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850121636578066432
author Xuxu Zheng
Peng Bao
Lin Qi
Chen Tian
Huawei Shen
author_facet Xuxu Zheng
Peng Bao
Lin Qi
Chen Tian
Huawei Shen
author_sort Xuxu Zheng
collection DOAJ
description The ability to predict the popularity of online contents has important implications in a wide range of areas. The challenge of this problem comes from the inequality of the popularity of content and the numerous complex factors. Existing works fall into three main paradigms: feature-driven approaches, generative models, and methods based on deep learning, each with known strengths and limitations. In this article, we propose an end-to-end deep learning framework, called CST-Net, to combat the defects of existing methods. We first learn a low-dimensional embedding for each user based on historic interactions. Then, users are clustered into communities based on the learned user embeddings, and information cascades are represented as a series of episodes in the form of community interaction matrix. Afterwards, a convolutional architecture is applied to learn the representation of the entire information cascade. Finally, the extracted structural and temporal features are further combined to predict the incremental popularity. We validate the effectiveness of the proposed CST-Net by applying it on two different types of population-scale datasets, i.e., a microblogging dataset and an academic citation dataset. Experimental results demonstrate that the proposed CST-Net model consistently outperforms the existing competitive popularity prediction methods.
format Article
id doaj-art-470405d5beb7449fbdeff1fb9bcc08f5
institution OA Journals
issn 2376-5992
language English
publishDate 2025-06-01
publisher PeerJ Inc.
record_format Article
series PeerJ Computer Science
spelling doaj-art-470405d5beb7449fbdeff1fb9bcc08f52025-08-20T02:35:01ZengPeerJ Inc.PeerJ Computer Science2376-59922025-06-0111e293110.7717/peerj-cs.2931CST-Net: community-guided structural-temporal convolutional networks for popularity predictionXuxu Zheng0Peng Bao1Lin Qi2Chen Tian3Huawei Shen4University of Chinese Academy of Sciences, Beijing, ChinaBeijing Jiaotong University, Beijing, ChinaBeijing Jiaotong University, Beijing, ChinaBeijing Jiaotong University, Beijing, ChinaUniversity of Chinese Academy of Sciences, Beijing, ChinaThe ability to predict the popularity of online contents has important implications in a wide range of areas. The challenge of this problem comes from the inequality of the popularity of content and the numerous complex factors. Existing works fall into three main paradigms: feature-driven approaches, generative models, and methods based on deep learning, each with known strengths and limitations. In this article, we propose an end-to-end deep learning framework, called CST-Net, to combat the defects of existing methods. We first learn a low-dimensional embedding for each user based on historic interactions. Then, users are clustered into communities based on the learned user embeddings, and information cascades are represented as a series of episodes in the form of community interaction matrix. Afterwards, a convolutional architecture is applied to learn the representation of the entire information cascade. Finally, the extracted structural and temporal features are further combined to predict the incremental popularity. We validate the effectiveness of the proposed CST-Net by applying it on two different types of population-scale datasets, i.e., a microblogging dataset and an academic citation dataset. Experimental results demonstrate that the proposed CST-Net model consistently outperforms the existing competitive popularity prediction methods.https://peerj.com/articles/cs-2931.pdfInformation diffusionPopularity predictionSocial networkNeural networks
spellingShingle Xuxu Zheng
Peng Bao
Lin Qi
Chen Tian
Huawei Shen
CST-Net: community-guided structural-temporal convolutional networks for popularity prediction
PeerJ Computer Science
Information diffusion
Popularity prediction
Social network
Neural networks
title CST-Net: community-guided structural-temporal convolutional networks for popularity prediction
title_full CST-Net: community-guided structural-temporal convolutional networks for popularity prediction
title_fullStr CST-Net: community-guided structural-temporal convolutional networks for popularity prediction
title_full_unstemmed CST-Net: community-guided structural-temporal convolutional networks for popularity prediction
title_short CST-Net: community-guided structural-temporal convolutional networks for popularity prediction
title_sort cst net community guided structural temporal convolutional networks for popularity prediction
topic Information diffusion
Popularity prediction
Social network
Neural networks
url https://peerj.com/articles/cs-2931.pdf
work_keys_str_mv AT xuxuzheng cstnetcommunityguidedstructuraltemporalconvolutionalnetworksforpopularityprediction
AT pengbao cstnetcommunityguidedstructuraltemporalconvolutionalnetworksforpopularityprediction
AT linqi cstnetcommunityguidedstructuraltemporalconvolutionalnetworksforpopularityprediction
AT chentian cstnetcommunityguidedstructuraltemporalconvolutionalnetworksforpopularityprediction
AT huaweishen cstnetcommunityguidedstructuraltemporalconvolutionalnetworksforpopularityprediction