Entity

Time filter

Source Type

Berwyn, Pennsylvania, United States

Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 200.00K | Year: 2011

This exploratory project (EMERG) is aimed at developing the capability to predict emerging topics in science from a highly detailed global model of the scientific literature. Identification of emergent opportunities in science is a central issue in academia and practice. Applications range from a simple understanding of the broader context in which individual research is conducted to the direction of research funds toward emerging topics. Previous studies of emergence have had the following shortcomings: they are retrospective (the area of emergence is identified after the fact), narrowly defined (lacking the context of related scientific topics) and/or highly aggregated (field level rather than topic level).

The approach is based on a highly detailed global model of science, consisting of hundreds of thousands of micro-communities over a period of nine years (2000-2008). The average size of these micro-communities is 15 papers per year. Micro-communities are linked from year to year using co-citation methods. Some micro-communities are part of long thread-like structures while others may be isolated. At the micro-structure level, science appears to have a high level of discontinuity. The mixture of continuity and discontinuity makes it possible to see emergence at the topical level. A variety of indicators, some structural, some based on the micro-community contents (articles, authors, ages, etc.), and some based on full text analysis, are calculated for each micro-community. The hypothesis is that several, if not all, of the proposed indicators will correlate with emergence.

To test this hypothesis, a data from research funding agencies and foundations that identified emergent micro-communities will be collected, together with identifying and tracking the key articles responsible for emergence in those areas. This history will be compared with the results of indicators from the model of science. If successful, the indicators can be applied to a current (rather than retrospective) model, suggesting the particular current micro-communities in science that are emerging or that are likely to emerge in the next year or two.

This project provides a completely new method for developing useful knowledge about the micro-structure and dynamics of science and technology from literature databases, whether of scientific literature, patent, grant, or web resources. This work has the potential to transform the way the structure and dynamics of science and technology are understood, and to impact conduct and management of research at the scientists, students, general public and policy maker levels. Project results will be disseminated via web site (http://www.mapofscience.com/emerg.html) and publications.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: STAR Metrics | Award Amount: 55.45K | Year: 2015

Understanding the outcomes associated with R&D funding requires accurate linking of funding with the papers produced by the funded work. This proposal develops a methodology to accurately link grants with topics, an important challenge for the development of a rigorous, quantitative understanding and analysis of science policy. The research will provide methods to accurately and consistently identify coherent research areas and systematically link those topic areas to research funding and research output, such as scientific publications, using text and other features.

The ability to accurately link grants and topics in a consistent way would be an important advance in the usefulness of STAR METRICS data. In addition to developing a methodology, this project also establishes the accuracy of existing NIH grant-to-article linkage data, and develop grant-to-article linkage data for non-NIH grants -- data which are not readily available. More accurate grant-article and grant-topic linkages will facilitate other research and be important to the STAR METRICS platform.

Discover hidden collaborations