tag:blogger.com,1999:blog-17889588.post1023884279541501421..comments2024-03-13T07:14:55.283+01:00Comments on chem-bla-ics: QSPR modeling with signaturesEgon Willighagenhttp://www.blogger.com/profile/07470952136305035540noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-17889588.post-50585552739002001482010-06-20T18:12:42.829+02:002010-06-20T18:12:42.829+02:00Yes, this model is predicting boiling points. It s...Yes, this model is predicting boiling points. It seems to work. The paper linked to in the blog actually discusses that many other theoretical indices can be written as signature-derived descriptors. For example, boiling point already quite well correlates to the number of atoms, which you can derived from the height=0 signatures.Egon Willighagenhttps://www.blogger.com/profile/07470952136305035540noreply@blogger.comtag:blogger.com,1999:blog-17889588.post-46194946210832299432010-06-20T17:57:36.280+02:002010-06-20T17:57:36.280+02:00Interesting, but I am not clear on what the model ...Interesting, but I am not clear on what the model is actually doing. You are predicting boiling points, is that right?<br /><br />Would extended valency sequences - like signatures - be likely to predict physical features like this?gilleainhttps://www.blogger.com/profile/14491887080861584059noreply@blogger.com