نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشیار گروه مدیریت اطلاعات، موسسه استنادی و پایش علم و فناوری جهان اسلام (ISC)، شیراز، ایران
2 دانشگاه شهید چمران اهواز
3 استادیار گروه مدیریت اطلاعات، مؤسسه استنادی و پایش علم و فناوری جهان اسلام (ISC)، شیراز، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Purpose: The present study addresses the gap in evaluating scientific impact beyond mere citation counts by introducing a network-based approach for assessing structural features of citation networks. Emphasis is placed on "Depth" and "Breadth" within Ego-Centric Citation Networks (ECCN). The investigation differentiates articles that foster deep, convergent knowledge clusters, often foundational or paradigm-shifting, from those exerting broad, diffuse influence across interdisciplinary domains. Attention is directed toward highly cited Iranian computer scientists identified via the Essential Science Indicators (ESI), thereby offering insights into citation patterns within a developing scientific context such as Iran. Such an orientation not only underscores the contributions of Iranian researchers to global computer science but also elucidates distinctive structural patterns shaped by national research policies, collaboration strategies, and access to international networks. The principal aim is to analyze the distribution of absolute and relative depth and breadth indices across these networks, thereby facilitating a more nuanced assessment of scientific impact. This methodology resonates with broader scientometric critiques that advocate structural analysis to augment count-based metrics, as citation enumeration alone may obscure distinctions between enduring and transient influence. Ultimately, the framework developed herein guides Iranian science policy by identifying high-impact works that propel innovation, bolster emerging paradigms, and facilitate cross-disciplinary diffusion, thereby supporting resource allocation, faculty promotions, and strategic research planning.
Methodology: The research employs a quantitative-analytical design based on network scientometrics with emphasis on ego-centric network analysis. The sample comprises 219 highly cited articles by 45 leading Iranian computer science researchers, as identified by the ESI. Selection relied on their ranking in the top 1% of citations in the field over a decade. Data were extracted from the Web of Science Core Collection (WOSCC) through web scraping with the MiMFa Scraper tool in Java. Processing was performed using MiMFa RAVAR DataLab in C# for network computations, SPSS V.26 for statistics, and Gephi for visualization. Validity was tested by comparing algorithms to manual checks. Deterministic algorithms and standardized WOSCC data ensured reliability.
Findings: Results show clear patterns in citation networks of highly cited Iranian computer science articles. In depth, Heidari et al.'s (2020) "Slime Mould Algorithm" topped with 1,781 citations, absolute depth of 2,228 (1,315 linkers and 913 secondary confirmers), and relative depth of 1.25. This indicates a dense, convergent network. Next came Maddah-Ali's (2014) "Fundamental Limits of Caching" (absolute depth: 1,340; relative: 1.34) and Sadollah et al.'s (2013) "Mine Blast Algorithm" (absolute: 960; relative: 1.43). Recent articles like Heidari et al.'s (2019) chaotic moth-flame optimizer reached the highest relative depth (1.44) with only 212 citations. This indicates rapid convergence in specialized groups. Articles from 2018 to 2021 had higher relative depth due to normalization and quick adoption in niche fields. Scatter plots showed a positive correlation between citations and absolute depth. Articles with more than 700 citations had depths exceeding 1,000, underscoring their role in complex knowledge clusters. Visualizations of networks, like Shakarami et al. (2020), revealed deep structures with strong links, stressing machine learning's integrative role. For breadth, Rahmani et al.'s (2021) global stroke burden study led with 2,013 citations, an absolute breadth of 1,257, and a relative breadth of 0.62. This reflects diffuse impact in interdisciplinary health informatics. Hatamlou's (2016) "Multi-Verse Optimizer" followed (absolute: 419; relative: 0.42). GBD-related articles exhibited high relative breadth (0.90–0.98), indicating sparse networks of cross-disciplinary citations. Heidari's (2020) article showed low relative breadth (0.14), confirming convergence. Distribution revealed 20% of articles with relative breadth under 0.2 (dense networks), 45% between 0.2 and 0.5 (semi-dense) and 35% over 0.5 (sparse). More recent and interdisciplinary articles had higher breadth. Scatter plots confirmed a positive correlation between citations and absolute breadth. Outliers are marked as foundational works that draw diverse, independent citations. Visualizations like those of Asghari et al. (2019) on IoT highlighted broad influence in dispersed settings.
Conclusion: The research shows that integrating depth and breadth in ego-centric citation networks creates a robust framework for qualitative impact assessment in computer science. It surpasses traditional metrics by uncovering structural details. Depth signals cohesive paradigm-shaping effects, while breadth highlights widespread innovative spread. Iranian articles display a dual pattern: dense networks in algorithmic advances and sparse ones in interdisciplinary uses. These findings stress the need for Iranian science policies to use such metrics for balanced assessment. This encourages foundational and emerging research. Limitations involve dependence on WOSCC coverage and biases in recent articles. Future work should incorporate other databases, international comparisons, and multi-layer networks to improve generalizability. The method facilitates efficient resource allocation and enhances innovation within Iran's scientific ecosystem.
کلیدواژهها [English]
)