tag:blogger.com,1999:blog-684443317148892945.post2451200084561156602..comments2024-02-29T23:54:39.092-08:00Comments on RDKit: Finding Related Documents in ChEMBLgreg landrumhttp://www.blogger.com/profile/10263150365422242369noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-684443317148892945.post-33533271290296090152013-12-11T02:08:48.049-08:002013-12-11T02:08:48.049-08:001) Good point about the mutual referencing thing. ...1) Good point about the mutual referencing thing. I had missed that. The limitation to one year was due to the performance of the search. I have this allergy to waiting more than an hour for results. ;-) Hopefully using chemfp will help with that.<br />2) It's no fun if you spoil my surprises! Seriously: it's is a much simpler analysis than SEA, but hopefully still useful.<br />greg landrumhttps://www.blogger.com/profile/10263150365422242369noreply@blogger.comtag:blogger.com,1999:blog-684443317148892945.post-74278945799717761882013-12-11T01:58:12.689-08:002013-12-11T01:58:12.689-08:00Really great post. Just a couple of things:
1) In ...Really great post. Just a couple of things:<br />1) In our current implementation, the limiting condition is that two papers have to cite/reference each other in order to proceed to molecule/structure similarity assesement. In your example, you looked at papers from 2012 only, which makes it unlikely to find papers that cite others in that set.<br />2) Your elegant cross-doc similarity snippet could be extended to a target level and form the basis of an 'Open SEA' implementation (hint, hint).<br /> <br />Anonymoushttps://www.blogger.com/profile/18414967060522745603noreply@blogger.com