Exploring dynamic research interest and academic influence for scientific collaborator recommendation

Kong, Xiangjie; Jiang, Huizhen; Wang, Wei; Bekele, Teshome Megersa; Xu, Zhenzhen; Wang, Meng

doi:10.1007/s11192-017-2485-9

Cited by 57 publications

(38 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most recommendation methods focus on providing items which are similar to the items liked or rated highly by the target user. Kong et al [23] exploited the similarity between scholars such as the dynamic research interests and academic influences for the collaborators recommendation. Recently, the technology of network representation has been widely used for solving various tasks by learning the representations of vertices in network with low dimensional vectors [24], i.e., designing different recommender systems based on the vector similarity.…”

Section: Serendipitous Recommendation Approachesmentioning

confidence: 99%

To Your Surprise: Identifying Serendipitous Collaborators

Wan

Yuan

Xia

et al. 2021

IEEE Trans. Big Data

View full text Add to dashboard Cite

Scientific collaboration has become a universal phenomenon in recent years. Meanwhile, scholars tend to hunt for surprising collaborators for broadening their horizons. Serendipity initially denotes the fortunate discovery. Although a lot of literature is available on the topic of serendipity, little research has investigated serendipity in scientific collaborations. The objective of this paper is to identify serendipitous scientific collaborators of target scholars based on their collaboration data. First, we induce the definition of serendipitous scientific collaborators by three components, which are relevance, unexpectedness, and value, respectively. They are quantified as three intuitive indices corresponding to the network proximity, topic diversity, and collaborator influence, respectively. Second, we propose a classification model, called RUVMod, to classify all collaborators based on the analysis of three indices in definition. The serendipitous collaborator has lower network proximity, higher topic diversity and higher influence than his/her target scholar relatively. Finally, we cluster all collaborators via Self Organizing Maps and identify the serendipitous collaborator class according to the classes divided in our RUVMod. We apply our definition to the scientific collaborators extracted from DBLP dataset. The evaluation from the serendipity-based metrics suggests that RUVMod is effective in identifying serendipitous scientific collaborators.

show abstract

Section: Serendipitous Recommendation Approachesmentioning

confidence: 99%

To Your Surprise: Identifying Serendipitous Collaborators

Wan

Yuan

Xia

et al. 2021

IEEE Trans. Big Data

View full text Add to dashboard Cite

show abstract

“…KullbackLeibler divergence is used. Also, reference [109] used generative probabilistic and similarity models to propose Beneficial Collaborator Recommendation (BCR). Also, reference [77] used generative language models for group finding task from heterogeneous document repository.…”

Section: Generative Probabilistic Modelsmentioning

confidence: 99%

Expert Finding Systems: A Systematic Review

et al. 2019

View full text Add to dashboard Cite

The data overload problem and the specific nature of the experts’ knowledge can hinder many users from finding experts with the expertise they required. There are several expert finding systems, which aim to solve the data overload problem and often recommend experts who can fulfil the users’ information needs. This study conducted a Systematic Literature Review on the state-of-the-art expert finding systems and expertise seeking studies published between 2010 and 2019. We used a systematic process to select ninety-six articles, consisting of 57 journals, 34 conference proceedings, three book chapters, and one thesis. This study analyses the domains of expert finding systems, expertise sources, methods, and datasets. It also discusses the differences between expertise retrieval and seeking. Moreover, it identifies the contextual factors that have been combined into expert finding systems. Finally, it identifies five gaps in expert finding systems for future research. This review indicated that ≈65% of expert finding systems are used in the academic domain. This review forms a basis for future expert finding systems research.

show abstract

“…They combined the content-based and random-walk based methods for collaborators recommendation. Kong's works (Kong et al, 2017(Kong et al, , 2016 exploited the dynamic research interests, academic influences and publication contents of scholars for collaborators recommendation.…”

Section: Random Walk-based Recommendationmentioning

confidence: 99%

Peer Review #2 of "A serendipity-biased Deepwalk for collaborators recommendation (v0.1)"

2019

View full text Add to dashboard Cite

Scientific collaboration has become a common behaviour in academia. Various recommendation strategies have been designed to provide relevant collaborators for the target scholars. However, scholars no longer satisfy with the acquainted collaborators recommendation since it may narrow their horizons. Serendipity in recommender system has attracted increasing attentions from researchers in recent years. Serendipity traditionally denotes the faculty of making surprising discoveries. The unexpected and valuable scientific discoveries in science such as X-rays and penicillin may be attributed to serendipity. In this paper, we design a novel recommender system to provide serendipitous scientific collaborators, which learns the serendipity-biased vector representation of each node in co-author network. We first introduce the definition of serendipitous collaborators from three components of serendipity: relevance, unexpectedness, and value, respectively. Then we improve the transition probability of random walk in DeepWalk, and propose a serendipity-biased DeepWalk, called Seren2vec. The walker jumps to the next neighbor node with the proportional probability of edge weight in co-author network. Meanwhile, the edge weight is determined by the three indices in definition. Finally, Top-N serendipitous collaborators are generated based on the cosine similarity between scholar vectors. We conduct extensive experiments on DBLP data set for validating our recommendation performance, and the evaluations from serendipity-based metrics show that Seren2vec outperforms other baseline methods without much loss of recommendation accuracy.

show abstract

Exploring dynamic research interest and academic influence for scientific collaborator recommendation

Cited by 57 publications

References 42 publications

To Your Surprise: Identifying Serendipitous Collaborators

To Your Surprise: Identifying Serendipitous Collaborators

Expert Finding Systems: A Systematic Review

Peer Review #2 of "A serendipity-biased Deepwalk for collaborators recommendation (v0.1)"

Contact Info

Product

Resources

About