Authors:
(1) Hamid Reza Saeidnia, Department of Information Science and Knowledge Studies, Tarbiat Modares University, Tehran, Islamic Republic of Iran;
(2) Elaheh Hosseini, Department of Information Science and Knowledge Studies, Faculty of Psychology and Educational Sciences, Alzahra University, Tehran, Islamic Republic of Iran;
(3) Shadi Abdoli, Department of Information Science, Université de Montreal, Montreal, Canada
(4) Marcel Ausloos, School of Business, University of Leicester, Leicester, UK and Bucharest University of Economic Studies, Bucharest, Romania.
Table of Links
RQ 4: Future of Scientometrics, Webometrics, and Bibliometrics with AI
RQ 5: Ethical Considerations of Scientometrics, Webometrics, and Bibliometrics with AI
Conclusion, Limitations, and References
Discussion
The above hopefully rather complete, at least extensive, literature survey allows a critical assessment of the state of AI in informatics science.
First, the findings in Table 1 have significant implications for scientometrics. They highlight the potential benefits and strategies for utilizing artificial intelligence (AI) capabilities in scientometrics analyses. The mentioned studies clearly demonstrate that AI can enhance the accuracy and efficiency of data collection and analysis in scientometrics [21, 22, 32, 33]. By automating various tasks, AI algorithms can reduce human errors and biases, ensuring more reliable and consistent results. This enhanced accuracy and efficiency save time and resources, allowing researchers to focus on high-level analyses and interpretations.
AI-based citation analysis methods, author disambiguation techniques, and predictive models showcased in the mentioned studies provide researchers with powerful tools for improving data collection and analysis in scientometrics [22, 24, 31, 34]. AI algorithms can effectively identify citation patterns, analyze the impact of scientific publications, and predict research trends. These capabilities enable researchers to gain deeper insights into the scientific landscape and make informed decisions.
Traditional citation counts have limitations in measuring research impact. However, the studies demonstrate that AI-based metrics can provide more comprehensive and accurate measures of research impact [25, 29]. By considering various factors beyond citations, such as social media mentions, downloads, and collaborations, AI algorithms can provide a more holistic view of the impact of scientific publications.
AI techniques showcased in the studies can analyze scientific literature to identify emerging research areas and patterns of scientific collaborations [28, 30]. This enables researchers to stay updated with the latest trends, discover new knowledge domains, and foster collaborations with relevant stakeholders.
AI-based peer review systems, as highlighted in one of the studies, can enhance the efficiency and objectivity of the peer review process [27, 57]. By automating parts of the review process, AI can ensure the publication of high-quality research, reduce biases, and provide faster feedback to authors. This improves the overall quality of scientometrics analyses and accelerates the dissemination of scientific knowledge.
Another study demonstrates that AI can assist in detecting instances of scientific misconduct, such as plagiarism and data fabrication [55]. By analyzing large volumes of data and comparing it against established standards, AI algorithms can identify potential cases of misconduct, ensuring the integrity of scientometrics analyses [17, 54, 55].
In summary, the findings in Table 1 demonstrate that AI has the potential to revolutionize techniques and approaches of scientometrics. AI capabilities improve the accuracy, efficiency, and reliability of data collection, analysis, and assessment of research impact. They enable the identification of emerging research areas, collaboration networks, and instances of scientific misconduct. Ultimately, these findings contribute to the advancement of scientometrics research, improving the quality, accessibility, and overall understanding of the scientific landscape.
Table 2 presents studies that demonstrate the potential benefits and strategies for utilizing artificial intelligence (AI) capabilities in webometrics [9, 36-44, 46, 58-66]. The findings in this table have significant implications for webometrics, as they highlight how AI can enhance various aspects of the field.
Indeed, the studies mentioned in Table 2 showcase that AI can improve data collection and analysis in webometrics, and how. In particular, AI algorithms can automate the process of gathering web data, such as web links, page content, and user behavior. This automation not only saves time and effort but also ensures the collection of larger and more diverse datasets, leading to more comprehensive webometric analyses.
AI techniques, such as machine learning and network analysis, are employed in the studies to improve web link analysis in webometrics [9, 43]. These techniques enable researchers to identify influential websites, web pages, and online communities [42, 59]. AI algorithms can analyze the structure and dynamics of web links, providing insights into the connectivity and impact of web resources [39, 42, 58, 59].
AI algorithms can analyze web content to extract relevant information and identify trends in webometrics [41, 62, 64, 65]. Natural language processing (NLP) techniques can be employed to automatically extract keywords, topics, and sentiments from web pages [40, 41, 62-65]. This automated analysis enhances the efficiency and accuracy of webometric studies, enabling researchers to gain insights into web-based information dissemination and trends [40, 41].
AI-based metrics and algorithms can provide advanced web impact assessment in webometrics [46, 60]. Beyond traditional link counts, AI algorithms can consider factors such as user behavior, social media mentions, and content engagement to measure the impact of web resources [37, 46, 60]. This comprehensive assessment helps researchers and organizations understand the reach and influence of web content [37, 46].
Web usage mining refers to the analysis of user behavior on the web. AI techniques, such as machine learning and data mining, can be employed to analyze user interactions, navigation paths, and preferences on websites. This analysis helps researchers understand user behavior patterns, improve web design, and enhance user experience.
AI algorithms can improve the efficiency and effectiveness of web crawling and data extraction in webometrics. These algorithms can automatically navigate through web pages, extract relevant data, and filter out irrelevant or duplicate information. This automation streamlines the data collection process, enabling researchers to analyze larger volumes of web data.
In a nutshell, let it be mentioned that the findings in Table 2 demonstrate that AI has the potential to significantly enhance webometrics. By improving data collection, web link analysis, content analysis, impact assessment, web usage mining, and data extraction, AI algorithms empower researchers to conduct more comprehensive and accurate webometric analyses. These advancements contribute to a deeper understanding of web-based information dissemination, user behavior, and the impact of web resources.
Thirdly, Table 3 presents studies that demonstrate the potential benefits and strategies for utilizing artificial intelligence (AI) capabilities in bibliometrics [21, 22, 24, 28, 30-34, 47-51, 53- 56, 67-72]. The findings in this table have significant implications for bibliometrics, as they highlight how AI can enhance various aspects of the field.
AI algorithms can improve publication analysis in bibliometrics [21, 22, 32, 33, 67]. By automatically extracting metadata from scientific publications, such as author names, affiliations, citations, and keywords, AI techniques can streamline the data collection process and improve accuracy [21, 22, 32, 33, 49, 67]. This automation allows researchers to analyze larger volumes of publications, facilitating comprehensive bibliometric analyses [21, 22, 32, 33, 67].
AI techniques can enhance citation analysis in bibliometrics. AI algorithms can automatically identify and analyze citation patterns, such as co-citation and bibliographic coupling [22, 24, 28, 31, 34, 68]. These algorithms can also identify citation networks and clusters, providing insights into the relationships among scientific publications [22, 24, 28, 31, 34, 68]. This analysis helps researchers understand the influence and impact of scholarly work [24, 31, 34].
AI algorithms can aid in author disambiguation, a critical task in bibliometrics [28, 30, 70- 72]. By analyzing various factors, such as author names, affiliations, and publication history, AI techniques can accurately identify and disambiguate authors with similar names [30, 72]. This disambiguation ensures accurate attribution of scholarly work and improves the reliability of bibliometric analyses [28, 30, 71, 72].
AI techniques, such as machine learning and data mining, can be employed to develop predictive models in bibliometrics [50, 51, 55, 56]. These models can forecast future publication trends, identify emerging research areas, and predict research impact [50, 51, 54, 55]. By analyzing patterns and relationships in large bibliographic datasets, AI algorithms can provide valuable insights into the future direction of scientific research [54-56].
AI algorithms can analyze collaboration networks among researchers in bibliometrics. By analyzing co-authorship patterns, affiliations, and research collaborations, AI techniques can identify influential researchers, research groups, and institutions. This analysis not only helps researchers understand the dynamics of collaboration but should be expected to foster interdisciplinary research, beside more usual links.
AI techniques can enhance research evaluation in bibliometrics. By considering various factors beyond traditional citation counts, such as social media mentions, downloads, and media coverage, AI algorithms can provide more comprehensive metrics for evaluating research impact. This comprehensive evaluation helps researchers, institutions, and funding agencies make informed decisions and allocate resources effectively.
Furthermore, improved or specifically written AI algorithms can assist in detecting instances of scientific misconduct, and prove plagiarism and data fabrication.
Concisely, the findings in Table 3 demonstrate that AI has the potential to significantly enhance bibliometrics. By improving publication analysis, citation analysis, author disambiguation, predictive models, collaboration analysis, and research evaluation, AI algorithms empower researchers to conduct more comprehensive and accurate bibliometric analyses. These advancements contribute to a deeper understanding of scholarly communication, research impact, and collaboration dynamics in the scientific community.
This paper is available on arxiv under CC BY 4.0 DEED license.