Timeline for Trouble implementing logistic regression
Current License: CC BY-SA 3.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 18, 2018 at 21:24 | history | reopened |
Nick Cox Juho Kokkala whuber♦ |
||
| Jan 18, 2018 at 15:50 | review | Reopen votes | |||
| Jan 18, 2018 at 21:24 | |||||
| Jan 18, 2018 at 15:36 | comment | added | Nick Cox | I disagree with the vote to close here. This is essentially about what logistic regression does and doesn't do and what to do instead. It's not focused on code or programming. The use of Stata in my answer is just to confirm the OP's calculations (their code being invisible here) . | |
| Jan 18, 2018 at 12:23 | history | closed |
Michael R. Chernick AdamO jbowman Peter Flom |
Not suitable for this site | |
| Jan 17, 2018 at 20:07 | comment | added | AdamO | @dfqupwndguerrillamailde not a linear term. You can play with quadratic or piecewise constant, but since it is a deterministic trend none will make much sense. | |
| Jan 17, 2018 at 20:02 | comment | added | dfqupwndguerrillamailde | @AdamO what model you do you reckon I should go with? | |
| Jan 17, 2018 at 19:43 | comment | added | AdamO | What is the trouble? You have a balanced quadratic trend in risk, so the prediction using a linear term is a flat response. The implementation was a success. The model choice was poor. | |
| Jan 17, 2018 at 18:52 | review | Close votes | |||
| Jan 18, 2018 at 12:23 | |||||
| Jan 17, 2018 at 18:49 | comment | added | whuber♦ | You don't have 1440 observations: you have three, of which two are incomplete! They consist of an interval ending at 480, the interval [481,1088], and an interval beginning at 1089. There's not a whole lot anyone can legitimately infer from such a small amount of data. What is the purpose of your app? Do you in practice have much more information? | |
| Jan 17, 2018 at 18:41 | history | edited | Nick Cox | CC BY-SA 3.0 |
edited body
|
| Jan 17, 2018 at 18:39 | answer | added | Nick Cox | timeline score: 4 | |
| Jan 17, 2018 at 18:30 | comment | added | Todd D | Your data has a striking pattern. What question are you trying to address with your regression? You may have a situation where logistic regression doesn't help much given the obvious pattern in your data. | |
| Jan 17, 2018 at 18:12 | review | First posts | |||
| Jan 17, 2018 at 18:35 | |||||
| Jan 17, 2018 at 18:12 | history | asked | dfqupwndguerrillamailde | CC BY-SA 3.0 |