Showing papers by "Ronald Rousseau published in 2014"

Measuring Scholarly Impact

Abstract: Large networks contain plentiful information about the organization of a system. The challenge is to extract useful information buried in the structure of myriad nodes and links. Therefore, powerful tools for simplifying and highlighting important structures in networks are essential for comprehending their organization. Such tools are called community-detection methods and they are designed to identify strongly intraconnected modules that often correspond to important functional units. Here we describe one such method, known as the map equation, and its accompanying algorithms for finding, evaluating, and visualizing the modular organization of networks. The map equation framework is very flexible and can identify two-level, multi-level, and overlapping organization in weighted, directed, and multiplex networks with its search algorithm Infomap. Because the map equation framework operates on the flow induced by the links of a network, it naturally captures flow of ideas and citation flow, and is therefore well-suited for analysis of bibliometric networks.

Measuring Scholarly Impact: Methods and Practice

Abstract: This book is an authoritative handbook of current topics, technologies and methodological approaches that may be used for the study of scholarly impact. The included methods cover a range of fields such as statistical sciences, scientific visualization, network analysis, text mining, and information retrieval. The techniques and tools enable researchers to investigate metric phenomena and to assess scholarly impact in new ways. Each chapter offers an introduction to the selected topic and outlines how the topic, technology or methodological approach may be applied to metrics-related research. Comprehensive and up-to-date, Measuring Scholarly Impact: Methods and Practice is designed for researchers and scholars interested in informetrics, scientometrics, and text mining. The hands-on perspective is also beneficial to advanced-level students in fields from computer science and statistics to information science.

Recommending research collaborations using link prediction and random forest classifiers

Abstract: We introduce a method to predict or recommend high-potential future (i.e., not yet realized) collaborations. The proposed method is based on a combination of link prediction and machine learning techniques. First, a weighted co-authorship network is constructed. We calculate scores for each node pair according to different measures called predictors. The resulting scores can be interpreted as indicative of the likelihood of future linkage for the given node pair. To determine the relative merit of each predictor, we train a random forest classifier on older data. The same classifier can then generate predictions for newer data. The top predictions are treated as recommendations for future collaboration. We apply the technique to research collaborations between cities in Africa, the Middle East and South-Asia, focusing on the topics of malaria and tuberculosis. Results show that the method yields accurate recommendations. Moreover, the method can be used to determine the relative strengths of each predictor.

Institution name disambiguation for research assessment

Abstract: Research evaluation is a necessity for management of academic units (scientists, research groups, departments, institutes, universities) and for government decision making in science and technology. Yet, wrong conclusions may be drawn due to errors in assignments of authors to institutions. To improve existing techniques of institution name disambiguation (IND) based on word similarity or editing distance, a rule-based algorithm is proposed in this study. One-to-many relationships between an institution and many variant names under which it is referred to in bylines of publications are recognized with the aid of statistical methods and specific rules. The performance of the rule based IND algorithm is evaluated on large datasets in four fields. These experimental results demonstrate that the precision of the algorithm is high. Yet, recall should be improved.

Scientists' referencing (mis)behavior revealed by the dissemination network of referencing errors

Abstract: This study illustrates scientists' referencing (mis)behavior by structuring the dissemination network of referencing errors. The sample set consists of 16,622 referencing errors of a highly cited paper published by Laemmli, UK in Nature in 1970. Dissemination networks of thirteen types of volume-page double errors and one type of page-only error are constructed and analyzed. Focusing on papers which carry the same volume-page double error, or the same page error, the citing-cited relationship between any two of them was identified and author bylines were compared to find common author(s). Our investigation results in three disseminating routes of referencing errors. Route 1: Citing a paper and copying its reference; Route 2: Copying a reference from another paper but without citing this paper; Route 3: Copying references from an earlier paper published by the author himself (herself) without rechecking the accuracy of the reference. The first two routes reflect scientists' referencing misbehavior while the third calls attention to self-copying of references.

Citation analysis and the development of science: A case study using articles by some Nobel prize winners

Abstract: Using citation data of articles written by some Nobel Prize winners in physics, we show that concave, convex, and straight curves represent different types of interactions between old ideas and new insights. These cases illustrate different diffusion characteristics of academic knowledge, depending on the nature of the knowledge in the new publications. This work adds to the study of the development of science and links this development to citation analysis.

A note on the interpolated or real-valued h-index with a generalization for fractional counting

Abstract: Purpose – The purpose of this paper is to extend the h-index framework to the case that articles are counted fractionally. Design/methodology/approach – Three restrictions related to the standard h-index are explained: as the standard h-index is a natural number it is a rather coarse indicator; if a scientist has published a relatively small number of publications then the h-index is completely determined by the number of publications; the standard h-index cannot be applied if publications are counted fractionally, or when magnitude values smaller than one occur. Findings – We recall solutions we proposed in earlier publications regarding the first two problems (the use of the interpolated h-index and of the pseudo h-index) and add a new proposal to solve the third problem. The relation between the recently introduced window/field-normalized h-type index (hwf-index) and the interpolated h-index is described. A real-world example proves the feasibility of this proposal. Research limitations/implications – ...

Increase in numbers and proportions of review articles in Tropical Medicine, Infectious Diseases, and oncology

Abstract: This article examines whether the absolute and relative numbers of reviews are increasing in the following three subfields of medical sciences: Tropical Medicine, Infectious Diseases, and Oncology. It further examines if reviews are cited more frequently than are “normal” articles. All research questions are answered affirmatively: The absolute as well as the relative numbers of reviews in these three subfields are indeed increasing. In addition, reviews in these fields are cited more frequently than are normal articles: about 70% more often than are “normal” articles in Infectious Diseases and Oncology and about 50% more often in Tropical Medicine. The article discusses possible reasons for this increase and concludes that medical journals should strive to achieve an optimal balance between review papers and original articles.

An approach for efficient online identification of the top-k percent most cited documents in large sets of Web of Science documents

Abstract: An approach for efficient online identification of the top-k percent most cited documents in large sets of Web of Science documents

Digital publishing and China's core scientific journals: a position paper

Abstract: In this position paper we discuss the current status of the core scientific journals in China. Based on discussions of journals' relation to a small group of full-text database providers, open access publishing and copyright problems, we conclude that China's digital publishing industry is not yet in a healthy state and some key issues related to revenue, digital piracy and copyright must be solved.

Describing the development of molecular research in the context of nervous system diseases using year-based h-cores

Abstract: This article introduces year-based h-indices as a tool to produce easy-to-use research overviews. Based on PubMed data, a general framework is constructed to study the development of molecular research in the context of nervous system diseases. It is shown which molecular substances are the centre of attention and which have passed their peak. Year-based h-indices have the following interesting features for representing changes in a field or discipline: flexibility, sensitivity to dynamic changes, fluctuation detection and trend detection.

Assessment of expertise overlap between an expert panel and research groups

Abstract: Discipline-specific research evaluation exercises are typically carried out by committees of peers, expert panels. Currently, there are no available methods that can measure overlap in expertise between a panel and the units of assessment. This research in progress paper explores a bibliometric approach to determining the overlap of expertise, using the 2010 research evaluation of nine physics research groups of the University of Antwerp as a test case. Overlay maps were applied to visualize to what extent the groups and panel members publish in different Web of Science subject categories. There seems to be a moderate disparity between the panel’s and the groups’ expertise. The panel was not as diverse as the groups that needed to be assessed. Future research will focus on journal level overlay maps, similarity testing, and a comparison with other disciplines.

A basic indicator for relative contributions and a remarkable difference between a ratio of averages and an average of ratios

Abstract: Egghe, L., & Rousseau, R. (2014). A basic indicator for relative contributions and a remarkable difference between a ratio of averages and an average of ratios. Malaysian Journal of Library & Information Science, Vol.19, no. 1: 17-22. This article has been inspired by the activity index and problems in its proper understanding. It studies a basic relative contribution index which is placed in the context of countries’ publication contributions. Two versions are proposed: one being an average of ratios (AoR) and the other one a ratio of averages (RoA). A Lotkaian-Zipfian framework is used to model the two versions of the proposed indicator. A remarkable difference between the two approaches (RoA vs. AoR) is found when determining the fraction of units (countries) that have a value larger than one. This observation contributes to the understanding of the differences between these two approaches.

Calculating the Outgrow Index and Similar Structural Indicators: A simple Software Program for Visualizing Outcomes

Abstract: It is shown that structural indicators, such as the outgrow index, can also be used in the context of diffusion studies. We, moreover, provide a simple software program to calculate and visualize the results.

A Preliminary Study of the Relationship between the h-Index and Excess Citations / Étude préliminaire de la relation entre l'indice de Hirsch (indice-h) et les citations excédentaires

Abstract: This article presents a study of the average number of excess citations of papers in the h -core, denoted as e 2 / h and the ratio between the e -area and the h -area, denoted as e 2 / h 2 . Real-world citation data from different countries are studied. It is found that at the country level, a small set of publications generate a disproportionately large amount of citations. Although different countries have different e 2 / h 2 values in different fields, average e 2 / h 2 values are all above 1. The e 2 / h values vary widely between fields, reflecting the general citation density in these fields. For cumulative data e 2 / h 2 and e 2 / h values each converge quickly. Neither a shifted Zipf nor an exponential model could fit the data. Resume: Cet article presente une etude du nombre moyen de citations excedentaires d’articles dans le core- h , note e2 /h et le rapport entre la zone- e et la zone- h , note e2 /h2 . Les donnees etudiees sont des donnees de citations de pays du monde reel. On constate qu’au niveau des pays, un petit groupe de publications genere un montant disproportionne de citations. Bien que differents pays aient des valeurs e2 /h2 differentes dans differents domaines, les valeurs moyennes e2 /h2 sont toutes superieures a l’unite. Les valeurs e2 /h varient considerablement suivant les domaines, ce qui reflete la densite generale de citation dans ces domaines. Pour les donnees cumulatives, chacune des valeurs e2 /h2 et e2 /h converge rapidement. Aucun modele zipfien decale ou exponentiel ne correspond aux donnees.