| Přispěvatelé: |
Raan, A.F.J. van, Tijssen, R.J.W., Bochhove, C.A. van, Breimer, D.D., Frenken, K., Frenken, J.W.M., Leiden University |
| Popis: |
The research presented in this PhD thesis describes ways of identifying at an early-stage, 2-3 years after their publication, discoveries in science that are expected to have a major impact on science. Bibliographic information extracted from those scientific publications is analysed to select patterns that may identify such `scientific breakthroughs’. A distinction is made between different types of breakthroughs. A major methodological issue is the differentiation between discoveries that actually achieve a long-term major impact on science and those where the impact has not occurred within the first three years. These impacts are measured in terms of `citations’ from subsequent research publications.This thesis introduces the conceptual framework that is used to analyses four case studies, each study focussing on a specific well-documented breakthrough discovery. The citation impact patterns are analysed in a search for bibliographic markers that are characteristic for a breakthrough discovery. The significance of the various markers, one set of markers for each type of breakthrough, were tested in a large-scale validation study.This study resulted in several early-stage identification algorithms. Five of these algorithms were implemented as automated computerized search methods. The algorithms have proved able, retrospectively, to identify early-stage publications that present (potential) breakthrough discoveries. |
