Visualizing Co-Authorship Social Networks and Collaboration Recommendations With CNARe

Visualizing Co-Authorship Social Networks and Collaboration Recommendations With CNARe

Michele A. Brandão (Universidade Federal de Minas Gerais, Brazil), Matheus A. Diniz (Universidade Federal de Minas Gerais, Brazil), Guilherme A. de Sousa (Universidade Federal de Minas Gerais, Brazil) and Mirella M. Moro (Universidade Federal de Minas Gerais, Brazil)
DOI: 10.4018/978-1-5225-2814-2.ch011
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Studies have analyzed social networks considering a plethora of metrics for different goals, from improving e-learning to recommend people and things. Here, we focus on large-scale social networks defined by researchers and their common published articles, which form co-authorship social networks. Then, we introduce CNARe, an online tool that analyzes the networks and present recommendations of collaborations based on three different algorithms (Affin, CORALS and MVCWalker). Through visualizations and social networks metrics, CNARe also allows to investigate how the recommendations affect the co-authorship social networks, how researchers' networks are in a central and eagle-eye context, and how the strength of ties behaves in large co-authorship social networks. Furthermore, users can upload their own network in CNARe and make their own recommendation and social network analysis.
Chapter Preview
Top

1. Introduction

Social networks represent individuals and the interactions among them, and studying such networks allows to discover different social patterns (Ahmed & Chen, 2016; Brandão & Moro, 2017). For instance, Chang & Chin (2011) study factors that affect user intention to use a social network game, and Pettenati & Cigognini (2007) use social networks theories to elaborate new e-learning practices. Furthermore, the social networks features can also be used to improve the quality of recommendation algorithms, such as those for friends, music, books and collaborators (He & Chu, 2010; Tang, Hu & Liu, 2013).

Specifically, recommending collaborators is a specific type of people recommendation in which the main goal is to recommend a pair of individuals to collaborate in a determined context. For instance, Surian et al., (2011) extract information from Source forge1 and build a developer collaboration network. Then, they propose a new algorithm to recommend developers candidate to projects in Source forge. Likewise, Protasiewicz et al., (2016) propose an architecture to recommend reviewers to evaluate researchers’ proposals and publications. In this context, this chapter focuses on recommendation of co-authors by considering algorithms that use information available in co-authorship social networks. A co-authorship social network is a type of social network in which nodes are authors and edges represent that they have publications in common.

Advances in collaboration recommendation algorithms have shown the potential to improve researchers’ productivity and their groups through establishing new research connections (Brandão et al., 2013; Lopes et al., 2010; Xia et al., 2014). The recommendation strategies include analyses of the topological features from the co-authorship social networks, semantic properties of the relationship between researchers and math formalizations. Such algorithms provide as result a recommendation list with the top ranked researchers that may collaborate with another researcher.

Besides characteristics of the recommendation algorithms, another relevant aspect of a full system is the visualization of the recommendations results. Generally, the recommendations are presented in sorted lists (according to the recommendation function’s result). For instance, Confer (used in IJCAI-162) is a tool that uses recommendation approaches in order to help conference attendees to find talks and papers, to discover people with common interest and manage their time in the conference (Zhang, Bhardwaj & Karger, 2016). It presents the recommendations as a list, and the users can attribute a star to each recommended item. However, these lists are often not enough to understand how the result was defined or to verify the potential of the recommendations to improve the network as a whole.

Here, the authors present an online tool called CNARe (Co-authorship Networks Analysis and Recommendations) - the pronounce is scenery (de Sousa et al., 2015). CNARe helps researchers to choose collaborators through automatic recommendations, visualize recommendations, compare the results from different recommendation algorithms and analyze the impact of the recommended researchers in their current network. The tool implements three recommendation algorithms (Brandão et al., 2013; Lopes et al., 2010; Xia et al., 2014). CNARe also provides other visualizations, for example, comparing the relationship between two or more co-authorship networks from different institutions and analyzing the strength of the co-authorships classified as social link (weak, strong or bridges - a co-authorship that connects researchers from different communities) or random relationship.

After discussing related work on recommender systems and social networks visualizations, the contributions of this chapter are summarized as follows.

Complete Chapter List

Search this Book:
Reset