meta user
network profile
| bio | website | |
|---|---|---|
| location | ||
| age | ||
| visits | member for | 1 year, 10 months |
| seen | Apr 26 at 14:34 | |
| stats | profile views | 160 |
|
Mar 22 |
awarded | Popular Question |
|
Mar 14 |
awarded | Popular Question |
|
Mar 1 |
awarded | Necromancer |
|
Feb 4 |
revised |
Suggested books on spatial statistics added 1 characters in body |
|
Feb 4 |
revised |
Suggested books on spatial statistics Corrected the title |
|
Feb 4 |
suggested | suggested edit on Suggested books on spatial statistics |
|
Feb 4 |
revised |
Suggested books on spatial statistics added 7 characters in body |
|
Feb 4 |
answered | Suggested books on spatial statistics |
|
Feb 3 |
comment |
Enormous coefficients in logistic regression - what does it mean and what to do? @JamesStanley 1) I understand, but are you sure this is also a separation? I think it is just because that if the reference level is not reliable (big CI), the other levels will also have big CI because they refer to it. But separation would mean that the level would predict the response without any error, which is different thing?? 2) thanks for the trick with relevel, it is new to me! Great! |
|
Feb 3 |
comment |
Enormous coefficients in logistic regression - what does it mean and what to do? Thanks @JamesStanley, checked this but would be probably caused by interaction rok:kraj in combination with very little data in the ULKV level of kraj. This will also cause high coefficient of krajULKV, because it is intercept for unreliable coefficient for rok:krajULKV |
|
Feb 3 |
comment |
Enormous coefficients in logistic regression - what does it mean and what to do? @zbicyclist, no, HKK level is quite frequent - and it wasn't me who drop it! This is done automatically in the model! Just the first level is dropped. The ULKV level of kraj is least frequent. Thanks. |
|
Feb 3 |
comment |
WinBUGS truncated normal distribution @DanielC Your error will be probably caused by different line of code than this one... |
|
Feb 3 |
comment |
WinBUGS truncated normal distribution The function djl.dnorm.trunc looks suspicious to me. Why don't you use the dnorm(..)I(..) here also? |
|
Feb 3 |
revised |
WinBUGS truncated normal distribution corrected typos, formatting |
|
Feb 3 |
answered | WinBUGS truncated normal distribution |
|
Feb 3 |
suggested | suggested edit on WinBUGS truncated normal distribution |
|
Feb 3 |
comment |
Suggested books on spatial statistics what area are you in? I have plenty of references on spatial statistics in ecology... but it might or might not be of use for you, depends on the area |
|
Feb 3 |
comment |
Enormous coefficients in logistic regression - what does it mean and what to do? James, one more question. Do you think that those model with quasi-complete separation would perform well in cross-validation or not? Because I tried it and the cross-validation seems to perform well (if I interpret it correctly). |
|
Feb 3 |
revised |
Intepretation of crossvalidation result - cv.glm() added 311 characters in body |
|
Feb 2 |
comment |
Enormous coefficients in logistic regression - what does it mean and what to do? Thank you James |
