Timeline for About calculating log-likelihood with zeroes
Current License: CC BY-SA 3.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 27, 2015 at 18:01 | comment | added | whuber♦ | The problem is universal to the blind use of black-box optimization algorithms, to be sure. It is important to code a function that can be computed for all values satisfying the constraints. For many algorithms, it is important that the function not fail for values that do not satisfy the constraints, too. | |
| Oct 27, 2015 at 15:48 | comment | added | Glen_b | Commenters weren't suggesting you "delete data". They were suggesting the parameter values should be discarded; they couldn't produce the data you have. | |
| Oct 27, 2015 at 14:56 | answer | added | Bill Woessner | timeline score: 4 | |
| Oct 27, 2015 at 14:30 | comment | added | mmh | @user777 I think the problem is universal to all optimization algorithms. I tried eight different solvers. | |
| Oct 27, 2015 at 14:25 | comment | added | mmh | The optimization algorithm decides which $a$ and $b$ to try. So, depending on the iteration different $x_{i}$'s are impossible. I think deleting data might introduce some problems. | |
| Oct 27, 2015 at 14:22 | history | edited | mmh | CC BY-SA 3.0 |
added 28 characters in body
|
| Oct 27, 2015 at 14:22 | comment | added | Stephan Kolassa | Well, if there are some $x_i<a$ or $x_i>b$, then you know that your $[a,b]$ can't be right. Your likelihood for such $a,b$ must be zero. Just discard them. | |
| Oct 27, 2015 at 14:22 | comment | added | Xi'an | These are impossible values so the likelihood is zero and the log likelihood minus infinity. | |
| Oct 27, 2015 at 14:21 | comment | added | mmh | Sorry there was a mistake in the text. It is now corrected. The problem appears with $x_{i} < a$ and $x_{i} > b$ where the density is zero. | |
| Oct 27, 2015 at 14:17 | history | asked | mmh | CC BY-SA 3.0 |