Timeline for Fit a multinomial model using proportions instead of raw data
Current License: CC BY-SA 3.0
32 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 5, 2019 at 6:02 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Mar 31, 2019 at 15:00 | history | tweeted | twitter.com/StackStats/status/1112369101538766848 | ||
| Dec 28, 2018 at 18:03 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jul 26, 2017 at 15:53 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
| Jun 6, 2017 at 2:19 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
| Jun 6, 2017 at 1:40 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
| Apr 21, 2017 at 8:12 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
deleted 19 characters in body; edited tags
|
| Mar 10, 2017 at 14:00 | history | edited | Agus Camacho | CC BY-SA 3.0 |
added 791 characters in body
|
| Mar 10, 2017 at 12:03 | history | edited | Agus Camacho | CC BY-SA 3.0 |
added 41 characters in body
|
| Mar 9, 2017 at 13:49 | history | edited | Agus Camacho | CC BY-SA 3.0 |
added 274 characters in body
|
| Mar 6, 2017 at 20:03 | comment | added | user78229 | This question seems like a near duplicate of this question ... stats.stackexchange.com/questions/254429/… | |
| Mar 6, 2017 at 19:40 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
deleted 3163 characters in body
|
| Mar 6, 2017 at 19:39 | comment | added | kjetil b halvorsen♦ | OK, see it! will edit to correct this. The pdf about mnlogit I linked also lists a coupe other available packages for multinomial logit. | |
| Mar 6, 2017 at 19:37 | comment | added | Agus Camacho | Dear Kjetil, i am very thankful for your help with this. Just three things, one is that is probably better to answer the question in the answers section. currenlty appears to me in the end of my own question. The second thing is that the dataset i created was a fake one, just for discussion purposes. The same behavior appears with my real dataset and probably any dataset. Finally, the mnlogit is not available for R version 3.3.2. | |
| Mar 4, 2017 at 14:07 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
added 3167 characters in body
|
| Mar 4, 2017 at 13:40 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
deleted 16 characters in body
|
| Mar 3, 2017 at 14:05 | history | edited | Agus Camacho | CC BY-SA 3.0 |
added 827 characters in body
|
| Mar 3, 2017 at 13:59 | comment | added | Agus Camacho | Added a reproducible example for a better discussion. Contrary to what was suggested here, relativizing the predictors reduces size effects and increased the p values. | |
| Mar 2, 2017 at 19:24 | history | edited | Agus Camacho | CC BY-SA 3.0 |
deleted 301 characters in body
|
| Mar 2, 2017 at 19:20 | vote | accept | Agus Camacho | ||
| Mar 3, 2017 at 13:56 | |||||
| Mar 2, 2017 at 16:37 | answer | added | kjetil b halvorsen♦ | timeline score: 1 | |
| Mar 2, 2017 at 16:27 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
added 690 characters in body
|
| Mar 2, 2017 at 15:22 | comment | added | kjetil b halvorsen♦ | This information should really be edited into the original post, you might get more interest and more answers that way! But, to answer question: NO NO, you will loose a lot of information, and the variance structure will be wrong. Use a function for multinomial logistic regression, with your original count vector, and maybe use correction for overdispersion. | |
| Mar 2, 2017 at 14:47 | comment | added | Agus Camacho | Thanks. i want to fit a non ordered model because the proportional odds ratio premise, necessary for using an ordered multinomial response does not hold for my dataset. That part I have previously solved. I obtain my observations from a hunting videogame in which the players may kill the prey, hurt it or fail, and the prey varies in speed and the size of vital body parts. I have thousands of observations. The only thing i need to know is whether or not i can use proportions instead of natural numbers as predictors in a non ordered multinomial model and why. | |
| Mar 2, 2017 at 14:23 | comment | added | kjetil b halvorsen♦ | Thanks, but what you described is still not very clear! You describe an ordered response: (success, partial success, fail) is clearly ordered! but then you explicitely want a non-ordered multinomial model. Why? If you can describe your context, your practical application, that is, successs with what? fail, in what way)? ... how you obtain the data, sample size, ... maybe we can understand! | |
| Mar 2, 2017 at 14:19 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
deleted 20 characters in body
|
| Mar 2, 2017 at 12:44 | history | edited | Agus Camacho | CC BY-SA 3.0 |
clarified the question
|
| Feb 25, 2017 at 23:39 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
added 1 character in body; edited title
|
| Feb 25, 2017 at 23:38 | comment | added | kjetil b halvorsen♦ | The question is not very clear, but ... first the response "success, partial success, fail" seems to be ordered, so you could try ordinal logistic regression. That would use the data mpre effectively than multinomial logistic regression. And, you need to use the counts, not the proportions. Using the proportions is throwing away a lot of information! | |
| Feb 24, 2017 at 13:14 | comment | added | kjetil b halvorsen♦ | Can you be more specific what data you have? Multinomial model is for count data, and what you have (tail length/body length) do not seem like that! Maybe beta regression or some generalization thereof? | |
| Feb 24, 2017 at 13:12 | history | edited | kjetil b halvorsen♦ | CC BY-SA 3.0 |
deleted 18 characters in body
|
| Feb 24, 2017 at 12:55 | history | asked | Agus Camacho | CC BY-SA 3.0 |