Timeline for Fitting a curve best practice
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 9, 2014 at 16:40 | comment | added | Ben Kuhn | @ganeshreddy, it's unclear from your comment, but if you don't have any data points in your training set with no experiment being run, and no idea how the $c_i$ getting added for each experiment relates to the $c$ that gets added when you have no experiment, I don't think there's much hope for modeling the no-experiment case. | |
| Dec 9, 2014 at 16:28 | comment | added | gbh. | Well the mixed model approach makes sense I agree but remember that finally we just want y = ax^2 + bx + c, for given x,a,b,c and thats it, we have no info on the experiment and such... | |
| Dec 9, 2014 at 16:13 | vote | accept | gbh. | ||
| Dec 9, 2014 at 13:25 | comment | added | Tim | You have a small sample for that, but the logic of linear mixed models would fit here: you would consider experiments as a random variable $N(0, \sigma)$, i.e. as a random sample of different possible conditions that could be encountered (so $c_i$ in @BenKuhn model would be a random effect). Notice also that you expect two things that rule out each other: (a) to include the impact of the experiment into the model, (b) not to include impact of the experiment in the model. | |
| Dec 8, 2014 at 23:11 | comment | added | Ben Kuhn | Do you have any tuples $(y, x)$ where no experiment is being run? | |
| Dec 8, 2014 at 22:35 | comment | added | gbh. | No, no. no and no :) In general the idea is to have a holistic model without the notion of experiments. | |
| Dec 8, 2014 at 22:19 | comment | added | Ben Kuhn | When predicting, will people also be using experiments that alter the value of $y$? If so, do you know anything about the distribution of the experiments they'll be using? Will they be using experiments that aren't in the training set at all? Do you know anything about the distribution of the $c_i$? | |
| Dec 8, 2014 at 22:15 | comment | added | gbh. | The idea is interesting and thanks for the answer, but somehow I don't like the idea of using experiment as a variable. Part of the reason is that while building a prediction, we will not have access to that. Is there another way to combine the data in an intelligent manner? Robust regression maybe? Also another issue is that some experiments may only have a single x,y pair | |
| Dec 8, 2014 at 22:11 | history | answered | Ben Kuhn | CC BY-SA 3.0 |