Timeline for How to perform a logistic regression for more than 2 response classes in R
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 7, 2015 at 13:59 | vote | accept | Juan Trujillo | ||
| S Oct 7, 2015 at 10:44 | history | suggested | Metariat | CC BY-SA 3.0 |
more clearance
|
| Oct 7, 2015 at 10:33 | review | Suggested edits | |||
| S Oct 7, 2015 at 10:44 | |||||
| Oct 7, 2015 at 10:30 | answer | added | Metariat | timeline score: 9 | |
| Oct 7, 2015 at 2:21 | comment | added | charles |
by definition logistic regression has two outcomes so you can (1) combine outcomes until you have two outcomes or (2) use an alternative method such as multinomial logistic regression available in multinom function from the nnet : ats.ucla.edu/stat/r/dae/mlogit.htm
|
|
| Oct 7, 2015 at 2:15 | comment | added | States.the.Obvious | Instead of logistic regression, you might want to consider linear discriminant analysis or quadratic discriminant analysis depending on the distribution of your data. | |
| Oct 6, 2015 at 23:41 | history | asked | Juan Trujillo | CC BY-SA 3.0 |