A Collaborators Recommendation Method Based on Multi-feature Fusion

Qi Yuan, Lujiao Shao, Xinyu Zhang, Xinrui Yu, Huiyue Sun, Jianghong Ma, Weizhi Meng, Xiao Zhi Gao, Haijun Zhang*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

This research introduces a new collaborators recommendation model based on multi-feature fusion. Specifically, we use a tree structure to integrate scholar information and extract content features from a scholar tree by using a Tree2vector-CLE model. Then, from the heterogeneous academic network, we extract the meta-path feature between scholars, which quantifies the similarity and co-operation potential between scholars from a multidimensional perspective. By combining content and meta-path features, we reconstruct a co-authorship network. Finally, we use the network representation learning method to represent the nodes in the reconstructed co-authorship network where the top-k collaborators are recommended for the target scholar with the random walk strategy controlled by the meta-path feature weighting. Experimental results on a real dataset demonstrate that our proposed method is effective in the task of collaborators recommendation.

Original languageEnglish
Title of host publicationNeural Computing for Advanced Applications
EditorsHaijun Zhang, Yuehui Chen, Xianghua Chu, Zhao Zhang, Tianyong Hao, Zhou Wu, Yimin Yang
PublisherSpringer
Publication date2022
Pages247-261
ISBN (Print)9789811961410
DOIs
Publication statusPublished - 2022
EventThird International Conference on Neural Computing for Advanced Applications - Jinan, China
Duration: 8 Jul 202210 Jul 2022
Conference number: 3

Conference

ConferenceThird International Conference on Neural Computing for Advanced Applications
Number3
Country/TerritoryChina
CityJinan
Period08/07/202210/07/2022
SeriesCommunications in Computer and Information Science
Volume1637 CCIS
ISSN1865-0929

Keywords

  • Collaborators recommendation
  • Meta-path features
  • Network representation learning
  • Scholars representation

Fingerprint

Dive into the research topics of 'A Collaborators Recommendation Method Based on Multi-feature Fusion'. Together they form a unique fingerprint.

Cite this