TLS: COLLABORATIVE RESEARCH: Tracking Scientific Innovation from Usage Data: Models and Tools to Support a Science of Science

该项目开发了一套工具，使投资于科学和工程的组织能够识别和预测创新研究的出现。这种能力将使组织能够有效地分配资源，以促进这些领域的快速有效的研究进程。需要几个关键属性：该工具应该能够实时运行，代表尽可能广泛的科学活动样本，并支持对分配资源的成本效益分析。智力价值：本研究旨在通过关注两个科学和方法论问题来支持此类工具的开发。首先，该项目研究科学活动早期指标的潜力，例如使用数据和搜索查询日志。其次，该项目旨在开发能够在此类早期指标的基础上实时识别和预测新兴趋势的模型。该项目利用了两个成熟项目的成果，即 MESUR 项目 (www.mesur.org) 和 Eigenfactor 项目 (www.eigenfactor.org)。在过去的两年中，MESUR 项目通过从世界上一些最重要的出版商、聚合商和大学联盟获取的超过 10 亿篇文章级使用事件，捕获了世界科学活动的重要样本。 Eigenfactor 项目展示了数学网络模型（参见 Google 的 PageRank）根据记录科学与工程研究集体历史的科学引文网格对学科和期刊进行排名的能力。对科学活动“流程”的预测已用于生成详细的科学活动地图，可以识别科学创新的潜在焦点。该项目扩展了特征因子模型，以包括 MESUR 的实际、实时科学活动指标。在此基础上，该项目开发了一套早期指标，可以在引文数据中看到科学创新的趋势之前实时检测科学创新的出现，并将这些指标与公共政策和决策联系起来。该项目还开发了解释和预测框架，将对个人行为的观察与科学创新等新兴集体现象联系起来。由于研究的重点是是否有可能开发出分析和预测工具来表明科学创新最有可能发生的原因、方式和地点，因此将利用现有的 eigenfactor.org 服务来生产免费的、可扩展的工具，对科学创新领域进行排名、分析、预测和图表。更广泛的影响：该研究项目提供免费的、可扩展的服务，以形成科学创新的“预警”系统，预计将导致公众更好地理解科学作为一个复杂的、动态的系统。此类服务应促进公众参与建立更加多样化、创新的研究格局，以应对 21 世纪的挑战。因此，这项工作应支持“更健康”的科学评估体系，通过承认塑造科学格局的更多样化的影响和贡献来促进创新。

This project develops a set of tools that allow organizations investing in Science and Engineering to identify and predict the emergence of innovative research. Such a capacity would permit organizations to efficiently allocate resources to stimulate rapid and effective research process in these areas. Several key attributes are needed: the tool should be able to operate in real-time, be representative of the widest possible sample of scientific activity, and support a cost-benefit analysis of allocated resources. Intellectual merit: This research aims to support the development of such tools by focusing on two scientific and methodological issues. First, the project studies the potential of early indicators of scientific activity such as usage data and search query logs. Second, the project aims to develop models that can, on the basis of such early indicators, identify and predict emerging trends in real-time.The project leverages the efforts of two well-established projects, namely the MESUR project (www.mesur.org) and the Eigenfactor project (www.eigenfactor.org). The MESUR project has, over the course of the past 2 years, captured a significant sample of the world?s scientific activity, via a collection of more than 1 billion article-level usage events acquired from some of the world's most significant publishers, aggregators and university consortia. The Eigenfactor project has demonstrated the power of mathematical network models (cf. Google's PageRank) to rank disciplines and journals according to the lattice work of scientific citations that records the collective history of S&E research. Predictions of the "flow" of scientific activity have been used to produce detailed maps of scientific activity that may identify potential foci of scientific innovation.This project expands the Eigenfactor models to include MESUR's indicators of actual, real-time scientific activity. On that basis the project develops a set of early indicators that can detect the emergence of scientific innovation in real-time - before such trends are visible in citation data - and relates these indicators to public policy and decision making. The project also develops explanatory and predictive frameworks that connect observations of individual behavior with emergent, collective phenomena such as scientific innovation. Since the focus of the research is whether it is possible to develop analytic and predictive tools that indicate why, how and where scientific innovation is most likely to occur, the existing eigenfactor.org services will be leveraged to produce freely available, expandable tools that rank, analyze, predict and chart areas of scientific innovation.Broader Impact: this research project produces freely available, expandable services to form an "early warning" system for scientific innovation that are expected to lead to a better public understanding of science as a complex, dynamic system. Such services should foster public participation in efforts to establish a more diverse, innovative research landscape that can meet the challenges of the 21st century.This work should thereby support a "healthier" system of scientific evaluation that fosters innovation by acknowledging a greater diversity of influences and contributions that shape the scientific landscape.