There are lots of "build a directed graph of bibliographic citations" tools.
If you can find a graph of relationships among papers using equations and that equation graph is better than bibliographic citation networks, that would be potentially useful to people who currently use citation-based graphs.
For example, a relatively "weak" correlation of math between two papers is the following. If paper1 has a+b and paper2 has a+b then place edge between those two papers. (That edge is weak because the use of the two equations might be a coincidence.)
For a slightly stronger correlation are there papers where a+b is used and there's a nearby citation to a bibliographic entry to a paper that also has the same a+b? That would be a stronger edge.
There are lots of "build a directed graph of bibliographic citations" tools. If you can find a graph of relationships among papers using equations and that equation graph is better than bibliographic citation networks, that would be potentially useful to people who currently use citation-based graphs.
For example, a relatively "weak" correlation of math between two papers is the following. If paper1 has
a+band paper2 hasa+bthen place edge between those two papers. (That edge is weak because the use of the two equations might be a coincidence.)For a slightly stronger correlation are there papers where
a+bis used and there's a nearby citation to a bibliographic entry to a paper that also has the samea+b? That would be a stronger edge.