First, thank you for the open-access and high-quality citation database! Apologies if this issue is in the wrong place.
I was looking to analyze the citation network as a directed acyclic graph, I wanted to ask a question about some of the records I came across where articles from the past seemed to cite articles from the future. There were two OCI ids I looked at for example:
Both of these "correctly" show the time-span as a negative value. Just wondering if this was an issue you were aware of, or anyone has attempted a topological sort or other guarantee that the citation network is a DAG. My specific use case was looking at out component (all research stemming from a given paper), but for many citations, cycles end up pulling in a huge amount of the citation graph. (The data I used were CSVs from the most recent figshare and analyzed them with graph-tool)
First, thank you for the open-access and high-quality citation database! Apologies if this issue is in the wrong place.
I was looking to analyze the citation network as a directed acyclic graph, I wanted to ask a question about some of the records I came across where articles from the past seemed to cite articles from the future. There were two OCI ids I looked at for example:
The first one, oci:0200309080736122422630004632858255908-020010000023600040701020604010800372427000106370001 Seems to have the citation flipped: 10.1002/0471264180.or016.01 is actually citing by 10.3987/com-04-s(p)8, not vice-versa.
For the second, oci:0200100020136122209090005070002-0200303090036252421342209000400010204 I actually wasn’t able to find a link in either direction between 10.3390/polym9040124 and 10.1021/cm9905702
Both of these "correctly" show the time-span as a negative value. Just wondering if this was an issue you were aware of, or anyone has attempted a topological sort or other guarantee that the citation network is a DAG. My specific use case was looking at out component (all research stemming from a given paper), but for many citations, cycles end up pulling in a huge amount of the citation graph. (The data I used were CSVs from the most recent figshare and analyzed them with graph-tool)