From owner-chemistry@ccl.net Fri Oct 29 09:21:00 2010 From: "Vincent Leroux vincent.leroux : loria.fr" To: CCL Subject: CCL: New Paper: Making Priors a Priority Message-Id: <-43034-101029075818-11483-o+NvvOKiLTeair8dvFt/Jg-$-server.ccl.net> X-Original-From: Vincent Leroux Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1 Date: Fri, 29 Oct 2010 13:58:08 +0200 MIME-Version: 1.0 Sent to CCL by: Vincent Leroux [vincent.leroux=loria.fr] Dear Matt Segall, You forgot to mention that you actually wrote this paper. Just imagine how this mailing list would look like if every scientist out there would advertise his/her last paper. Plus you dare call an "on-line community" part of the website of the commercial company you work for. By looking at the forum seems like "community" might be a little bit to big a word to describe it, but it is for sure a more adequate place for your post ;-) If it was a Science or Nature paper about computational chemistry, I would be happy to get the opportunity to discuss with the authors here, but well, even if your paper might be interesting we are still far from there... "How useful will a model be in making a decision in a practical context?". I want to know too! Two nice corporate keywords "making a decision" and "practical context" for the price of one: well done! Congratulations for "Making Priors a Priority". Awesome title! "To answer this requires an understanding of the prior probability distribution (the prior) and hence prevalence of negative outcomes due to the property being assessed.": sure this is right! May I add: "It also requires an understanding of the present probability distribution (the present) and hence actual significance of the property being assessed, and an understanding of the future probability (the prospect) that things will still make some kind of sense, provided that prior of present did". Oh wait, I did not read the paper yet, maybe it is in the conclusion already, just before the usual "These interesting questions will be investigated in future works" that could be called "Making Presents Presentable" and "Prospecting Prospects". Regards, VL Le 29/10/10 10:21, Matt Segall matt.segall{}optibrium.com a écrit : > Sent to CCL by: "Matt Segall" [matt.segall---optibrium.com] > A new paper, "Making Priors a Priority" addresses a key challenge the community > needs to address to get the most out of the predictive models we build. Just > published on-line in J. Comp. Aided Mol. Design (DOI 10.1007/s10822-010-9388- > 7), the paper discusses the question, "How useful will a model be in making a > decision in a practical context?" To answer this requires an understanding of > the prior probability distribution (the prior) and hence prevalence of > negative outcomes due to the property being assessed. > > The full paper can be downloaded from the Springer website at > http://www.springerlink.com/content/x15h763770611316/ or a free preprint is > available from the Optibrium on-line community at > http://www.optibrium.com/community/publications/110-making-priors-a-priority.> > >