Heterogeneous Academic Networks Research Articles

Metapath and attribute-based academic collaborator recommendation in heterogeneous academic networks

Metapath and attribute-based academic collaborator recommendation in heterogeneous academic networks

Anomalous citations detection in academic networks

Citation network analysis attracts increasing attention from disciplines of complex network analysis and science of science. One big challenge in this regard is that there are unreasonable citations in citation networks, i.e., cited papers are not relevant to the citing paper. Existing research on citation analysis has primarily concentrated on the contents and ignored the complex relations between academic entities. In this paper, we propose a novel research topic, that is, how to detect anomalous citations. To be specific, we first define anomalous citations and propose a unified framework, named ACTION, to detect anomalous citations in a heterogeneous academic network. ACTION is established based on non-negative matrix factorization and network representation learning, which considers not only the relevance of citation contents but also the relationships among academic entities including journals, papers, and authors. To evaluate the performance of ACTION, we construct three anomalous citation datasets. Experimental results demonstrate the effectiveness of the proposed method. Detecting anomalous citations carry profound significance for academic fairness.

AIRank: An algorithm on evaluating the academic influence of papers based on heterogeneous academic network

Evaluation of papers’ academic influence is a hot issue in the field of scientific research management. Academic big data provides a data treasure with the coexistence of different types of academic entities, which can be used to evaluate academic influence from a more macro and comprehensive perspective. Based on academic big data, a heterogeneous academic network composed of links within and between three types of academic entities (authors, papers and venues) is constructed. In addition, a new academic influence ranking algorithm, AIRank, is proposed to evaluate papers’ academic influence. Different from the existing academic influence ranking algorithms, AIRank has made innovations in the following two aspects. (1) AIRank distinguishes the influence transmission intensity between different node pairs. Different from the strategy of evenly distributing influence among different node pairs, AIRank quantifies the intensity of influence transmission between node pairs based on investigating the citation emotional attribute, semantic similarity and academic quality differences between node pairs. Based on the intensity characteristics, AIRank realises the distribution and transmission of influence among different node pairs. (2) AIRank incorporates the influence transmission from heterogeneous neighbours in evaluating papers’ influence. According to the academic influence of author nodes and venue nodes, AIRank fine-tunes the iteration formula of paper influence to obtain the ranking of papers under the joint influence of homogeneous and heterogeneous neighbours. Experimental results show that, compared with the ranking results based on citation frequency and PageRank algorithm, AIRank algorithm can produce more differentiated and reasonable academic influence ranking results.

Topic-sensitive expert finding based solely on heterogeneous academic networks

Identifying experts in a specific research field is an essential and practical task in academia and industry. Although some efforts have been attempted for expert finding, they rely on abundant text mining to reveal the target research topic accurately. Furthermore, it is time-consuming to collect and process these text materials for each candidate, and it is difficult to prepare a precise description for a topic that can cover all sub-topics. To this end, in this study, we develop a new expert finding model called TOpic-Sensitive representative Author identification (TOSA), to find representative authors in a specific research topic from a heterogeneous topic-insensitive academic network involving various topics, with a few well-known experts collected previously who are professional in the target topic. Specifically, we first learn the high-order relationships about authors in the heterogeneous academic network, which contain enough topic-related information. Next, an embedding space is found, where the collected experts are close to each other; in other words, the area they gather indicates the features about the collected authoritative experts. Finally, a prototype is calculated based on the collected experts in the embedding space, and we can rank other unlabeled authors by measuring the closeness between them and the prototype. Along this line, in addition to the collected experts in advance, more authoritative experts in the target topic will be identified. Extensive experiments on DBLP and ACM datasets demonstrate that the proposed TOSA model can identify experts who are professional in the query topic directly from a heterogeneous topic-insensitive academic network without text data processing, and its performance is significantly superior to the baseline models.

SentenceLDA- and ConNetClus-Based Heterogeneous Academic Network Analysis for Publication Ranking

Scientific papers published in journals or conferences, also considered academic publications, are the manifestation of scientific research achievements. Lots of scientific papers published in digital form bring new challenges for academic evaluation and information retrieval. Therefore, research on the ranking method of scientific papers is significant for the management and evaluation of academic resources. In this paper, we first identify internal and external factors for evaluating scientific papers and propose a publication ranking method based on an analysis of a heterogeneous academic network. We use four types of metadata (i.e., author, venue (journal or conference), topic, and title) as vertexes for creating the network; in there, the topics are trained by the SentenceLDA algorithm with the metadata of the abstract. We then use the Gibbs sampling method to create a heterogeneous academic network and apply the ConNetClus algorithm to calculate the probability value of publication ranking. To evaluate the significance of the method proposed in this paper, we compare the ranking results with BM25, PageRank, etc., and homogeneous networks in MAP and NDCG. As shown in our evaluation results, the performance of the method we propose in this paper is better than other baselines for ranking publications.

Graph Embedding for Scholar Recommendation in Academic Social Networks

The academic social networks (ASNs) play an important role in promoting scientific collaboration and innovation in academic society. Accompanying the tremendous growth of scholarly big data, finding suitable scholars on ASNs for collaboration has become more difficult. Different from friend recommendation in conventional social networks, scholar recommendation in ASNs usually involves different academic entities (e.g., scholars, scientific publications, and status updates) and various relationships (e.g., collaboration relationship between team members, citations, and co-authorships), which forms a complex heterogeneous academic network. Our goal is to recommend potential similar scholars for users in ASNs. In this article, we propose to design a graph embedding-based scholar recommendation system by leveraging academic auxiliary information. First, we construct enhanced ASNs by integrating two types of academic features extracted from scholars’ academic information with original network topology. Then, the refined feature representations of the scholars are obtained by a graph embedding framework, which helps the system measure the similarity between scholars based on their representation vectors. Finally, the system generates potential similar scholars for users in ASNs for the final recommendation. We evaluate the effectiveness of our model on five real-world datasets: SCHOLAT, Zhihu, APS, Yelp and Gowalla. The experimental results demonstrate that our model is effective and achieves promising improvements than the other competitive baselines.

Measuring academic entities\u2019 impact by content-based citation analysis in a heterogeneous academic network

Evaluating the impact of papers, researchers and venues objectively is of great significance to academia and beyond. This may help researchers, research organizations, and government agencies in various ways, such as helping researchers find valuable papers and authoritative venues and helping research organizations identify good researchers. A few studies find that rather than treating citations equally, differentiating them is a promising way for impact evaluation of academic entities. However, most of those methods are metadata-based only and do not consider contents of cited and citing papers; while a few content-based methods are not sophisticated, and further improvement is possible. In this paper, we study the citation relationships between entities by content-based approaches. Especially, an ensemble learning method is used to classify citations into different strength types, and a word-embedding based method is used to estimate topical similarity of the citing and cited papers. A heterogeneous network is constructed with the weighted citation links and several other features. Based on the heterogeneous network that consists of three types of entities, we apply an iterative PageRank-like method to rank the impact of papers, authors and venues at the same time through mutual reinforcement. Experiments are conducted on an ACL dataset, and the results demonstrate that our method greatly outperforms state-of-the art competitors in improving ranking effectiveness of papers, authors and venues, as well as in being robust against malicious manipulation of citations.

Collaboration prediction in heterogeneous academic network with dynamic structure and topic

Collaboration prediction in heterogeneous academic network with dynamic structure and topic

Heterogeneous academic network embedding based multivariate random-walk model for predicting scientific impact

The prediction of current scientific impact of papers and authors has been extensively studied to help researchers find valuable papers and recent research directions, also help policymakers make recruitment decisions or funding allocation. However, how to accurately evaluate the future impact of them, especially for new papers and young researchers, is the focus of scientific impact prediction research, and is less explored. Existing graph-based methods heavily depend on the global structure information of heterogeneous academic network and ignore the local structure information and text information, which may provide important clues to identify influential papers and authors with novel perspective. In this paper, we propose a hybrid model called ESMR to predict the future influence of papers and authors by mainly exploiting these information mentioned above. Specifically, we first put forward a novel network embedding-based model, which can capture not only the local structure information, but also the text information of papers into a unified embedding representation. Then, the future impact of papers and authors is mutually ranked by integrating the learned embedding representations into a multivariate random-walk model. Empirical results on two real datasets demonstrate that the proposed method significantly outperforms the existing state-of-the-art ranking methods.

Collaborative Filtering With Network Representation Learning for Citation Recommendation

Citation recommendation plays an important role in the context of scholarly big data, where finding relevant papers has become more difficult because of information overload. Applying traditional collaborative filtering (CF) to citation recommendation is challenging due to the cold start problem and the lack of paper ratings. To address these challenges, in this article, we propose a collaborative filtering with network representation learning framework for citation recommendation, namely CNCRec, which is a hybrid user-based CF considering both paper content and network topology. It aims at recommending citations in heterogeneous academic information networks. CNCRec creates the paper rating matrix based on attributed citation network representation learning, where the attributes are topics extracted from the paper text information. Meanwhile, the learned representations of attributed collaboration network is utilized to improve the selection of nearest neighbors. By harnessing the power of network representation learning, CNCRec is able to make full use of the whole citation network topology compared with previous context-aware network-based models. Extensive experiments on both DBLP and APS datasets show that the proposed method outperforms state-of-the-art methods in terms of precision, recall, and MRR (Mean Reciprocal Rank). Moreover, CNCRec can better solve the data sparsity problem compared with other CF-based baselines.

A Unified Link Prediction Framework for Predicting Arbitrary Relations in Heterogeneous Academic Networks

Most of the existing link prediction methods for heterogeneous academic networks can only predict one or two specific relation types rather than arbitrary relation types. Although several recently proposed methods have involved multi-relational prediction problems, they do not comprehensively consider the rich semantic or temporal information of heterogeneous academic networks. Considering that researchers may have diverse requirements for different types of academic resources, in this study, we propose a new unified link prediction framework (UniLPF) for arbitrary types of academic relations. First, a weighted and directed heterogeneous academic network containing rich academic objects and relations is constructed. Then, an automatic meta-path searching method is proposed to extract the meta-paths for arbitrary prediction tasks. Two meta-path based object similarity measures combining temporal information and content relevance are also proposed to measure the features of the meta-paths. Finally, a pervasive link prediction model is built, which can be embodied based on an arbitrarily specified prediction task and the corresponding meta-path features. Extensive experiments for predicting various relation types with practical significance are conducted on a large-scale Microsoft Academic dataset. The experimental results demonstrate that our proposed UniLPF framework can predict arbitrary specified academic relations, and outperforms the comparison methods in terms of F-measure, accuracy, AUC and ROC. In addition, the time scalability experiments prove that UniLPF also achieves good performance for predicting the academic relations over time.

Predicting Long-Term Scientific Impact Based on Multi-Field Feature Extraction

Nowadays, there have been many studies on evaluating the scientific impact of scholars. However, we still lack effective methods to predict long-term impact, especially 10 years in the future. Therefore, we propose a long-term scientific impact prediction model based on multi-field feature extraction. The workflow of our proposed model consists of feature engineering and model ensemble. In feature engineering, we extract attribute feature, time-series feature, and heterogeneous network feature based on three different fields. Moreover, when extracting heterogeneous network feature, we propose a scientific impact evaluation method based on heterogeneous academic network, which considers both the time of publication and author order factors. In the model ensemble, we adjust the basic model and noise model to the different training set to make full use of the information from the original dataset. The experiment results demonstrate that the proposed model can stably improve the accuracy of scholars' scientific impact prediction, and it also offers a prediction pattern for long-term prediction problem.

AIRank: Author Impact Ranking through Positions in Collaboration Networks

Citation is a universally acknowledged way for scientific impact evaluation. However, due to its easy manipulability, simply relying on citation cannot objectively reflect the actual impact of scholars. Instead of citation, we utilize the academic networks, in virtue of their available and abundant academic information, to evaluate the scientific impact of scholars in this paper. Through the collaboration among scholars in academic networks, we notice an interesting phenomenon that scholars in some special positions can access more kinds of information and connect researchers from different groups to promote the scientific collaborations. However, this important fact is generally ignored by the existing approaches. Motivated by the observations above, we propose the novel method AIRank to evaluate the scientific impact of scholars. Our method not only considers the impact of scholars through the mutual reinforcement process in heterogeneous academic networks, but also integrates the structural holes theory and information entropy theory to depict the benefit that scholars obtain via their positions in the network. The experimental results demonstrate the effectiveness of AIRank in evaluating the impact of scholars more comprehensively and finding more top ranking scholars with interdisciplinary nature.

Hete_MESE: Multi-Dimensional Community Detection Algorithm Based on Multiplex Network Extraction and Seed Expansion for Heterogeneous Information Networks

Most real-world information networks are heterogeneous, which contain multiple types of entities and relations between the entities. Large-scale and heterogeneity are the typical properties of heterogeneous information networks, and their community structures are often overlapping, complex, and diverse. The existing community detection algorithms without considering the above-mentioned properties may lead to low accuracy or high time complexity of community detection. In this paper, we study the multi-dimensional heterogeneous community detection problem for large-scale heterogeneous information networks with general topology structures. We define the concept of node-centric community and propose a multi-dimensional community detection algorithm referred as Hete_MESE. Specifically, we first specify one of the multiple entity types in heterogeneous information networks as community centric node type and extract the multiplex networks accordingly, and then, overlapping node-centric communities are detected based on the multiplex networks, which are considered as the seed communities to absorb the other entity types to generate the heterogeneous communities utilizing seed expansion. Taking heterogeneous academic network as an example, the effects of Hete_MESE are evaluated through extensive experiments, and the features of detected multi-dimensional academic communities are analyzed. The results demonstrate that Hete_MESE can accurately and effectively detect the meaningful overlapping and heterogeneous communities in heterogeneous information networks with general topologies from multiple dimensions. Moreover, Hete_MESE has linear time complexity, so it can be applied to large-scale heterogeneous information networks.

Exploring time factors in measuring the scientific impact of scholars

Taking advantage of the easy access to the rich and massive scholarly data, more and more researchers are focusing on the studies of analyzing and utilizing the scholarly big data. Among them, evaluating the scientific impact of scholars has significant importance. Measuring the scientific impact of scholars can not only provide basis for the applications of academic foundations and awards, but also shed light on the research directions for scholars. Currently, citation based methods and network based metrics are the most commonly used ways to evaluate the scientific impact. However, these approaches ignore several important facts, i.e. the dynamics of citations and the initial qualities of different articles. To alleviate the shortcomings of them, we propose a Time-aware Ranking algorithm (TRank) to evaluate the impact of scholars. Due to scholars' sustainable supreme concerns of academic innovations, the TRank algorithm gives more credits to the newly published scholarly papers as well as their references according to the representative time functions. Our method also combines the merits of random walk algorithms and heterogeneous network topology, i.e. the mutual influences among different scholarly entities in heterogeneous academic networks. To validate the suitable time function for TRank algorithm and explore its performance, we construct the experiments on two real datasets: (1) Digitial Bibliography and Library Project, and (2) American Physical Society. The experimental results demonstrate that our algorithm outperforms other methods in selecting outstanding scholars and the evaluation results on the overall impact of scholars.

Scientific Ranking over Heterogeneous Academic Hypernetwork

Ranking is an important way of retrieving authoritative papers from a large scientific literature database. Current state-of-the-art exploits the flat structure of the heterogeneous academic network to achieve a better ranking of scientific articles, however, ignores the multinomial nature of the multidimensional relationships between different types of academic entities. This paper proposes a novel mutual ranking algorithm based on the multinomial heterogeneous academic hypernetwork, which serves as a generalized model of a scientific literature database. The proposed algorithm is demonstrated effective through extensive evaluation against well-known IR metrics on a well-established benchmarking environment based on the ACL Anthology Network.

Query-dependent cross-domain ranking in heterogeneous network

Traditional learning-to-rank problem mainly focuses on one single type of objects. However, with the rapid growth of the Web 2.0, ranking over multiple interrelated and heterogeneous objects becomes a common situation, e.g., the heterogeneous academic network. In this scenario, one may have much training data for some type of objects (e.g. conferences) while only very few for the interested types of objects (e.g. authors). Thus, the two important questions are: (1) Given a networked data set, how could one borrow supervision from other types of objects in order to build an accurate ranking model for the interested objects with insufficient supervision? (2) If there are links between different objects, how can we exploit their relationships for improved ranking performance? In this work, we first propose a regularized framework called HCDRank to simultaneously minimize two loss functions related to these two domains. Then, we extend the approach by exploiting the link information between heterogeneous objects. We conduct a theoretical analysis to the proposed approach and derive its generalization bound to demonstrate how the two related domains could help each other in learning ranking functions. Experimental results on three different genres of data sets demonstrate the effectiveness of the proposed approaches.