Search citation statements
Paper Sections
Citation Types
Year Published
Publication Types
Relationship
Authors
Journals
Online social networks (OSNs) have become one of the most popular platforms where people communicate by sharing contents and personal information. The interactions performed by the users allow to identify the homophily between users and reveal the presence of several communities that could depend on several factors: such as the type of relationships (eg, colleagues and school mates) or to users' preferences (eg, users' interests or hobbies). A very important issue in this scenario is the necessary to characterize such communities by using known real properties or attributes about their members. In this article, we propose an approach that identifies the communities of users by exploiting several community detection algorithms. Afterward, for each user, we exploit decision trees to find a model that describes and distinguishes community affiliations based on known attributes of the members. The evaluation of our approach is derived from a real dataset which consists of the profile information, relationships, and interactions of 95 716 Facebook users. The experimental results show that the proposed approach is able to correctly recognize which attributes of the members properly characterize their corresponding community while ensuring a high level of accuracy (about 85%).
Online social networks (OSNs) have become one of the most popular platforms where people communicate by sharing contents and personal information. The interactions performed by the users allow to identify the homophily between users and reveal the presence of several communities that could depend on several factors: such as the type of relationships (eg, colleagues and school mates) or to users' preferences (eg, users' interests or hobbies). A very important issue in this scenario is the necessary to characterize such communities by using known real properties or attributes about their members. In this article, we propose an approach that identifies the communities of users by exploiting several community detection algorithms. Afterward, for each user, we exploit decision trees to find a model that describes and distinguishes community affiliations based on known attributes of the members. The evaluation of our approach is derived from a real dataset which consists of the profile information, relationships, and interactions of 95 716 Facebook users. The experimental results show that the proposed approach is able to correctly recognize which attributes of the members properly characterize their corresponding community while ensuring a high level of accuracy (about 85%).
Advisor-advisee relationship among scholars is important in the academia circle. It contains abundant information about the academic inheritance, advisor recommendation and the forming of research communities, etc. The advisor-advisee relationship is always hiding behind the co-author network, however, there are some challenges when mining this kind of relationship. This relationship is always changing with time, the size of labeled data is limited and the authors' name ambiguity, etc. Previous works are focused on various aspects, including the citation network, the publication network and the co-author network, etc. To our best knowledge, all of these works are focused on the whole network, and none of them considered the credit allocation of the authors in each paper. Therefore, the relationship mining results may be influenced greatly by some high degree nodes. In this paper, we proposed a new method to solve this problem with the scholar data in DBLP. The credit allocation of each author is calculated, and the co-author network of DBLP is cut into smaller networks based on the characteristic. Then, the advisor-advisee relationship among researchers is mined based on these smaller co-author network. The results show that, the accuracy of this model is about 62.5%, however, this is an unsupervised method, which could save the time of training model and will not be influenced by the uncompleted training data set.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.