Enable researchers to claim/identify multiple copies of their own publication that are located in various places like an author copy in their own datasource, a preprint in the arxiv datasource as well as a published version in a journal's datasource.
Approach
The issue for journal and author copy duplicates is straight forward: The journal version has an identifier (e.g. DOI) that the author can add to the version in their own datasource, since they can manipulate that at any time.
Somewhat more challenging is the issue of the preprint. Since the author in general does not have control over the preprint version of his work, located in the preprint server's datasource, the author can't add an identifier to the preprint, once it is issued by the journal. When searching for the publication the author copy and the journal version could be merged by the software client due to their common identifier. However, the preprint version can't be merged since it does not have the identifier.
How to merge the preprint?
Metadata extracted from the publication could be used to calculate a similarity measure between two publications to merge them in case they exceed a similarity threshold.
ORCiD identifiers could be used to create a more robust version of the aforementioned metadata extraction approach.
Goal
Enable researchers to claim/identify multiple copies of their own publication that are located in various places like an author copy in their own datasource, a preprint in the arxiv datasource as well as a published version in a journal's datasource.
Approach
The issue for journal and author copy duplicates is straight forward: The journal version has an identifier (e.g. DOI) that the author can add to the version in their own datasource, since they can manipulate that at any time. Somewhat more challenging is the issue of the preprint. Since the author in general does not have control over the preprint version of his work, located in the preprint server's datasource, the author can't add an identifier to the preprint, once it is issued by the journal. When searching for the publication the author copy and the journal version could be merged by the software client due to their common identifier. However, the preprint version can't be merged since it does not have the identifier.
How to merge the preprint?
Metadata extracted from the publication could be used to calculate a similarity measure between two publications to merge them in case they exceed a similarity threshold.
ORCiD identifiers could be used to create a more robust version of the aforementioned metadata extraction approach.