Timeline for Averaging model Coefficients in binary logistic regression
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 21, 2017 at 14:31 | comment | added | Sue | @hxd1011. Thank you for the advise. I would like to try averaging models to see how my results differ. would you know a simplified method of averaging I can try, or direct me to literature I can read? When I ran my models each with a different set of 75% of the data, I realize that the estimates consistently differ among the covariates but sometimes differ greatly between model. For example a variable having a coefficient of 0.123 in one model can have a coefficient of 0.132 in another model or 0.232 in another. Is this a sign or sampling errors in data collection? | |
| Mar 21, 2017 at 14:30 | history | edited | Haitao Du | CC BY-SA 3.0 |
added 110 characters in body
|
| Mar 21, 2017 at 14:21 | comment | added | Haitao Du | @HongOoi I am trying to say, linear model usually has high bias low variance in terms of bias variance trade off. Thanks for your comment I will revise it to make it more clear. | |
| Mar 21, 2017 at 14:19 | history | edited | Haitao Du | CC BY-SA 3.0 |
added 2 characters in body
|
| Mar 21, 2017 at 14:18 | comment | added | Hong Ooi | "high bias model like logistic regression" -- huh? | |
| Mar 21, 2017 at 14:11 | history | edited | Haitao Du | CC BY-SA 3.0 |
added 494 characters in body
|
| Mar 21, 2017 at 14:03 | history | answered | Haitao Du | CC BY-SA 3.0 |