The Metrics for Promising R&D Early Forecast
The paper is concerned to the new methodological approach for early prognosis of promising research and development (R&D) in the fields of science and technology. As it’s demonstrated with the texts’ corpus covering the known R&D stories from 1955 to 2014, the correct prognosis probability a...
Guardado en:
| Autor principal: | |
|---|---|
| Formato: | article |
| Lenguaje: | EN RU |
| Publicado: |
North-West institute of management of the Russian Presidential Academy of National Economy and Public Administration
2018
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/47033fbc49d7468794a868d8117bff9d |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | The paper is concerned to the new methodological approach for early prognosis of promising research and development (R&D) in the fields of science and technology. As it’s demonstrated with the texts’ corpus covering the known R&D stories from 1955 to 2014, the correct prognosis probability at the R&D life pre-investment cycle phase based on the traditional econometric methods is not high. The analogy is drawn to computer program text quality metrics based test approach and formally well-structured scientific & engineering texts. Making a start from this analogy, the four size-based quantitative metrics and four functional-based ones basing on lexical approach completed with ontological evolutionary approach to R&D texts’ corpus investigations are worked out. The relevant formulas are deduced to calculate the size-based metrics. The resulting values are interpreted form the point of view of promising R&D search and prognosis task. The key questions are described in details for source data formation to calculate more complex functional-based metrics using some lexical-graph R&D text models, to solve decomposition tasks and path search on graphs of terms collocations and co-words with the purpose of terminology evolution investigations, tautological definitions localization, and texts structure quality estimation. The source data necessary for the eight deduced formulas of author’s metrics calculation are rigorously specified. The non-linear pair correlation indexes are evaluated for every metric and known R&D historical result on the test text corpus. The probabilities of correct forecast with the eight metrics demonstrate good level of correlation with successful R&D stories. The ranges of resulting values for all the metrics are rigorously described and interpreted, their details of correlation indexes behavior and correct forecast probabilities are explained to support good decision regarding the most promising R&D choice and fulfill a purpose of investment activity at the early phase of R&D life cycle. As it’s demonstrated by the author the implementation of described mathematical approach based on the eight metrics results in higher probability of prognosis for better R&D choice and lets an investment manager to achieve the purpose of optimal funding in combination with other known methods. |
|---|