Abstract For many public research organizations, funding creation of science and maximizing scientific output is of central interest. Typically, when evaluating scientific production for funding, citations are utilized as a proxy, although these are severely influenced by factors beyond scientific impact. This study aims to investigate the consequences of the Matthew effect in citations, where prominent authors and prestigious journals receive more citations regardless of the scientific content of the publications. To this end, the study presents statistical models for the prediction of citations of papers based solely on observable characteristics available at the submission stage of a double-blind peer-review process. Combining classical linear models, generalized linear models and utilizing large-scale data sets on biomedical papers based on the PubMed database, the results demonstrate that it is possible to make fairly accurate predictions of citations using only observable characteristics of papers excluding information on authors and journals, thereby mitigating the Matthew effect. Thus, the outcomes have important implications for the field of scientometrics, providing a more objective method for citation prediction by relying on pre-publication variables that are immune to manipulation by authors and journals, thereby enhancing the objectivity of the evaluation process. Our approach is thus important for government agencies responsible for funding the creation of high-quality scientific content rather than perpetuating prestige.

Abstract In rapidly evolving scientific domains such as edge computing, timely identification of emerging research directions is increasingly critical. Preprints, as fast and openly accessible scholarly outputs, have long served as an important medium for the early dissemination of scholarly knowledge, particularly in high-velocity fields. However, their lack of peer review and heterogeneous quality present challenges for systematic analysis. This study examines edge computing preprints published between 2021 and 2023, retrieved from the Web of Science Preprint Citation Index. A hybrid topic modeling approach combining Latent Dirichlet Allocation (LDA) and Word2Vec-based semantic similarity was applied, along with an institutional credibility–based filtering mechanism to ensure corpus quality. The findings reveal three prominent thematic clusters: (1) resource management and task scheduling, focusing on optimizing computational infrastructure; (2) edge intelligence and federated learning, highlighting distributed AI applications; and (3) real-time analytics and low-latency processing, reflecting network efficiency concerns. These results align with prior research and suggest that preprints can complement traditional literature in capturing early research signals. The proposed framework demonstrates adaptability to dynamic, unstructured corpora and offers a scalable method for real-time knowledge monitoring in emerging technological fields.

Abstract Bibliometric analysis, also known as bibliometrics, has been conducted for several decades to evaluate scientific research based on data available on bibliographic platforms, such as the popular Web of Science or Scopus. Research papers which include bibliometric analysis typically ignore the problem of bibliographic data preprocessing, in particular its important aspect—data cleaning. Discussion of bibliographic data preprocessing in the literature is sparse and scattered; studies usually address selected single components of the entire endeavour. This study aims to fill the gap as a review article, extensively analysing the problem, presenting issues arising from the structure of bibliographic data, combining data from various sources, creating thesauri and conducting bibliometric analyses, also through the author’s own experience. A brief description of the most popular software dedicated to bibliometrics, such as BibExcel, Bibliometrix, CiteSpace, CitNetExplorer, SciMAT, Sci2 Tool, and VOSviewer, is also provided, highlighting the operations available in these applications for the preliminary processing of bibliographic data. The work allows us to draw the following conclusions. The task is more difficult and demanding than some authors suggest or unclearly claim has already been accomplished, without providing additional details. Data cleaning operations are carried out at various stages of preprocessing, sometimes repetitively, and the order in which they are performed may be significant as it determines the success or failure of the process, in particular when combining data from different sources. There is no software which allows automatic execution of the entire preprocessing procedure of bibliographic data. Moreover, manual work is inevitable at various stages of the process. The contribution of this work to the field of bibliometric analysis is expressed in the form of a methodological synthesis, which involves the holistic consideration of the discussed issue, enabling a more comprehensive understanding of it.

Abstract The parameter that captures the similarity among disciplinary categories is a key quantity of many measures of interdisciplinarity. This study evaluates the feasibility of using large language models to estimate this parameter rather than using traditional methods based on citational networks among disciplines. An experimental procedure tested the precision, agreement, resilience, robustness, and explainability of estimates from OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude. The experiment collected a sample of 228 similarity matrices among two disciplinary taxonomies, for a total of 16,200 sampled estimate values. The experiment concludes that Gemini reaches precise estimates, comparable to traditional methods. ChatGPT stands out only for its superior resilience when dealing with semantically trivial changes in how disciplines are described. Claude resulted in a balanced profile. While rarely in full agreement, all three models undertake the estimation task sufficiently well.

Abstract Understanding the benefits and challenges of international mobility programmes for both home and host countries is essential for evaluating them. This study provides a quantitative assessment of the Chinese government-sponsored international mobility programmes, the Chinese Scholarship Council (CSC) programmes, by examining their effects on home and host countries across disciplines from five key perspectives: output volume, impact, collaboration, funding, and involvement in security-sensitive research areas. Our findings suggest that, for the home country (China), the funding programmes enhance China’s research capabilities in alignment with its strategic development priorities. Additional evidence suggests that the funding programmes facilitate international collaboration and allow awardees to work with highly cited researchers. For host countries, the funding programmes promote collaboration with China and may support host researchers where funding might be scarce. Also, the funding programmes place little emphasis on dual-use research and have limited ties to the “Seven Sons of National Defence” universities. This finding addresses current concerns in some host countries regarding security challenges. The study provides evidence for policymakers to develop strategies that maximise the benefits of international mobility programmes and address associated challenges.