Scientific Methods Metadata for Provision of the Methods Reuse and Result Reproducibility.
|Author(s): || Skvortsov N. A., Briukhov D. O., Kalinichenko L. A., Kovalev D. Y., Stupnikov S. A.
|Published:||Proc. of the 15th Russian Conference on Digital Libraries RCDL'2013. -- CEUR Workshop Proceedings, 2013. – Vol. 1108. – P. 70--78. (In Russian)
| Data-intensive sciences are characterized by the
constantly growing needs for specific data analysis
methods intended for producing new knowledge related
to the investigated areas. Development of new data
analysis methods becomes a significant, inseparable part
of research. Digital preservation, reuse and
reproducibility of computer experiment results become
inherent attributes of scientific discovery. The paper
investigates metadata structure to be attached to the
processes specifying or implementing scientific data
analysis methods for their reuse and result
77reproducibility. Process components and data are
referred to the domain concepts and need to be supplied
with the information about data provenance and quality.
Specific test collections are needed to describe kinds of
cases in which methods must behave in an anticipated
way. Using the open myExperiment environment
organizing and providing access to the collection of
scientific workflows as an illustration, we demonstrate
how the extension of its metadata could have allowed to
organize the semantic search for methods relevant to a
problem, to verify interoperability, reusability and
reproducibility of processes implementing the methods.
[ Adobe PDF ]