Timeline for Understanding Marginalization of Uncertain Variables
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 16, 2019 at 14:54 | comment | added | GENIVI-LEARNER | well now it makes so much sense! You have no idea how much I appreciate your contribution to the answer as this was a cruicial step for me trying to understand Batch Bayesian optimization which is still unanswered. Could you please take a look at it and see if there is some thing you could contribute, no matter how small. | |
| Nov 16, 2019 at 9:30 | comment | added | rxFt20 | @GENIVI-LEARNER I updated the answer. I would suggest perhaps making a new post if you have further specific questions, since it will likely mean more eyes on the problem. Hope that helps. | |
| Nov 16, 2019 at 9:28 | history | edited | rxFt20 | CC BY-SA 4.0 |
Typo
|
| Nov 15, 2019 at 20:01 | comment | added | GENIVI-LEARNER | Also does the equation written above {product of the uncertain probabilities} and then the integration makes sense? | |
| Nov 15, 2019 at 19:49 | vote | accept | GENIVI-LEARNER | ||
| Nov 15, 2019 at 19:46 | comment | added | GENIVI-LEARNER | Also does integrating out same as taking "expected value"? | |
| Nov 15, 2019 at 19:45 | comment | added | GENIVI-LEARNER | Im glad you answered. I do have a clarification. First of all there are two sources of uncertainty, one is "uncertainty in measurement" and the other "uncertainty in prediction". So by integrating out "uncertainty in measurement" are we obtaining predictive distribution on the new point $z_{new}$ conditioned on the whole data {certain data + integrated out uncertain data} | |
| Nov 15, 2019 at 19:26 | history | answered | rxFt20 | CC BY-SA 4.0 |