Timeline for penalized regression applications in epidemiology
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 11, 2013 at 18:11 | vote | accept | user2300643 | ||
| Dec 11, 2013 at 15:50 | answer | added | Thomas Speidel | timeline score: 3 | |
| Dec 11, 2013 at 13:41 | comment | added | Frank Harrell | @user2300643 it wouldn't hurt to read my book or wait approx. 9 months for the 2nd edition. And I find DAGs help stimulate thought but are a bit overrated in their ability to lead us to a final conclusion. | |
| Dec 11, 2013 at 2:05 | comment | added | user2300643 | @EpiGrad I know of an epi professor who loves DAGs, which I think are useful for explaining confounding; however, I find it difficult when trying to use it to diagram synergistic effects, though I will admit that I do not utilize it enough to feel proficient. | |
| Dec 11, 2013 at 2:05 | comment | added | user2300643 | @Frank Harrell Thank you for clarifying penalized and shrinkage models. I am finding your course notes to be useful. I also found your book in Amazon.com. Would it behoove me to read it through? | |
| Dec 10, 2013 at 20:59 | comment | added | charles | (1) I have been meaning to read up on Cady-Allen type procedures. May be theoretically sounder than classic backwards selection (2) a priori model selection does seem to be strongly favored in the field. | |
| Dec 10, 2013 at 20:48 | comment | added | Fomite | @user2300643 Epidemiology as a field has been heavily influenced by several methodologists who are strongly against most selection algorithms. If you want a more systematic consideration of variable selection, consider directed acyclic graphs (DAGs)? A simple Google search will bring up a great many examples. | |
| Dec 10, 2013 at 20:31 | comment | added | Frank Harrell | You are mixing penalization with variable selection. lasso does that but this needn't be the case, and in general the best predictions are made using shrinkage of all parameters without setting any parameters to zero. | |
| Dec 10, 2013 at 20:17 | comment | added | user2300643 | Thanks, Charles. Sorry I was revising my comment about the CIs, but it appears I have deleted the comment. Given that a shrinkage model also includes bias and does not include SEs, then it would appear the shrinkage model would remain in the domain of prediction - for me anyway- as I wouldn't feel "confident" to extrapolate meaning from the coefficients... What are some other reliable feature selection methods than stepwise that might be useful for epidemiological studies? The populations I work with are indigenous and have little published available works. | |
| Dec 10, 2013 at 19:58 | comment | added | charles | In most textbooks there is a reference now to the LASSO or similar shrinkage methods. But the lack of consensus on how the estimates should be interpreted and CI estimated is a significant issue. Most references I see either don't mention this or relegate this to a brief aside (like the linked article). | |
| Dec 10, 2013 at 19:19 | comment | added | charles | Will this method still be of use if you have shrunken estimates without standard errors? You might be interested in: stats.stackexchange.com/questions/7225/… | |
| Dec 10, 2013 at 19:03 | history | asked | user2300643 | CC BY-SA 3.0 |