[目的/意义]运用深度学习技术,提出结合时间和空间特征的测度(速度、覆盖度和迂回度)方法,用于量化学者研究主题演化,从而为基于内容的学者评价提供量化依据。[方法/过程]提出三维指标框架,其中速度反映作者改变研究主题快慢的平均程度,覆盖度反映作者研究内容所覆盖的主题广度,迂回度反映作者研究路径的曲折性。使用微软学术数据集中计算机科学的作者进行实证研究,并考察学者研究主题演化的三维测度和学者学术影响力和生产力的关系。[结果/结论] 实证研究结果显示,覆盖度与总被引量和总发文量的关系为单调递减,这一特征说明聚焦于特定研究主题较为深入的作者,其发文量和影响力都较大。作者研究主题演化的"速度"和"迂回度"与总被引量、总发文量都存在先增加后减少的倒U型关系。所提出的多维度指标框架不仅可在理论上丰富科学计量学对于学者研究主题转移演化及其机制的理解,而且结合深度学习模型提出了问题的解决思路。
[Purpose/Significance] This paper proposes temporal-spatial measurements for research topic evolution of researchers, namely speed, volume, and circuitousness, with deep learning techniques. The measurements are adopted to quantify the research topic evolution of researchers and for quantitatively evaluate researchers based upon their research content. [Method/Process] Among the three indicators, "speed" indicated how fast an author shifts his/her research topics, "volume" implied the width of his/her research topics, and "circuitousness" illustrated how "tortuous" an author’s research topic evolution "path" looked like. This paper adopted authors of computer science recorded in the Microsoft Academic Graph (MAG) dataset and examined the three-dimensional measurement of the evolution of scholars’ research topics and the relationship between scientific impact and productivity of researchers. [Result/Conclusion] Empirical results show that the relationship between volume and total number of citations (total number of publications) is monotonically decreasing, which indicates that authors who focus on specific research topics more deeply have greater publications and influence. Speed and circuitousness of the author’s research topic evolution have an inverted U-shaped relationship (first increase and then decrease) with the total number of citations and the total number of publications. The multi-dimensional indicator framework proposed in this paper not only enriches scientometric understanding of scholars’ research topic shift evolution and its mechanism but also proposes potential solutions to the issues with deep learning-related models.
[1] JIA T, WANG D, SZYMANSKI B K. Quantifying patterns of research-interest evolution[J]. Nature human behaviour, 2017, 1(4): 0078.
[2] ZENG A, SHEN Z, ZHOU J, et al. Increasing trend of scientists to switch between topics[J]. Nature communications, 2019, 10(1): 1-11.
[3] 霍朝光, 张斌, 董克. 复杂网络视域下的学术行为预测研究述评: 选题, 合作与引用[J]. 情报理论与实践, 2021, 44(6): 180-188.
[4] TOUBIA O, BERGER J, ELIASHBERG J, et al. How quantifying the shape of stories predicts their success[J]. Proceedings of the National Academy of Sciences of the United States of America, 2021, 118(26): e2011695118.
[5] JEONG Y K, SONG M, DING Y. Content-based author co-citation analysis[J]. Journal of informetrics, 2014, 8(1): 197-211.
[6] ZHOU J, FAN Y. Quantifying the research diversification of physicists[J]. Journal of systems science and systems engineering, 2021, 30(6): 1-16.
[7] SINATRA R, DEVILE P, SZELL M, et al. A century of physics[J]. Nature physics, 2015, 11(10): 791-796.
[8] ZHANG C, BU Y, DING Y, et al. Understanding scientific collaboration: homophily, transitivity, and preferential attachment[J]. Journal of the Association for Information Science and Technology, 2018, 69(1): 72-86.
[9] LIU M, BU Y, CHEN C, et al. Pandemics are catalysts of scientific novelty: evidence from COVID-19 [J]. Journal of the Association for Information Science and Technology, 2022, 73(8):1065-1078.
[10] DEVLIN J, CHANG MW, LEE K, et al. Bert: pre-training of deep bidirectional transformers for language understanding[J]. arXiv preprint arXiv:1810.04805, 2018.
[11] SINHA A, SHEN Z, SONG Y, et al. An overview of Microsoft academic service (mas) and applications[C]// Proceedings of the 24th international conference on World Wide Web. Florence: ACM, 2015: 243-246.
[12] WANG K, SHEN Z, HUANG C Y, et al. Microsoft academic graph: when experts are not enough[J]. Quantitative science studies, 2020, 1(1): 396-413.
[13] ALSHEBLI B K, RAHWAN T, WOON W L. The preeminence of ethnic diversity in scientific collaboration[J]. Nature communications, 2018, 9(1): 1-10.
[14] BU Y, LI M, GU W, et al. Topic diversity: a discipline scheme-free diversity measurement for journals[J]. Journal of the Association for Information Science and Technology, 2021, 72(5): 523-539.
[15] GUPTA B, KARISIDDIPPA C. Collaboration and author productivity: a study with a new variable in Lotka’s law[J]. Scientometrics, 1999, 44(1): 129-134.
[16] LEYDESDORFF L. How are new citation-based journal indicators adding to the bibliometric toolbox?[J]. Journal of the American Society for Information Science and Technology, 2009, 60(7): 1327-1336.
[17] LEVITT J M, THELWALL M. Is multidisciplinary research more highly cited? a macrolevel study[J]. Journal of the American Society for Information Science and Technology, 2008, 59(12): 1973-1984.
[18] LIU L, DEHMAMY N, CHOWN J, et al. Understanding the onset of hot streaks across artistic, cultural, and scientific careers[J]. arXiv preprint arXiv:2103.01256, 2021.
[19] LEYDESDORFF L, RAFOLS I, MILOJEVIC S. Bridging the divide between qualitative and quantitative science studies[J]. Quantitative science studies, 2020, 1(3): 918-926.
[20] 付伟棠. 我国学术期刊同行评议研究综述[J]. 中国科技期刊研究, 2019, 30(8): 819-826.
[21] 步一, 许家伟, 黄文彬. 基于引文的科学文献定量评价:引文影响力指标评述[J]. 图书情报知识, 2021, 38(6): 47-59.
[22] WALTMAN L. A review of the literature on citation impact indicators[J]. Journal of informetrics, 2016, 10(2): 365-391.
[23] DING Y. Applying weighted PageRank to author citation networks[J]. Journal of the American Society for Information Science and Technology, 2011, 62(2): 236-245.
[24] 赵丹群. 文献计量范式下的科学知识图谱研究: 新进展与新挑战[J]. 情报学进展, 2020, 13: 354-380.
[25] 张琳, 刘冬东, 吕琦, 等. 论文学科交叉测度研究: 从全部引文到章节引文[J]. 情报学报, 2020, 39(5): 492-499.
