Timeline for What are the advantages of linear regression over quantile regression?
Current License: CC BY-SA 4.0
19 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| S Mar 15, 2019 at 19:58 | history | bounty ended | Peter Flom | ||
| S Mar 15, 2019 at 19:58 | history | notice removed | Peter Flom | ||
| Mar 15, 2019 at 19:58 | vote | accept | Peter Flom | ||
| Mar 15, 2019 at 15:38 | answer | added | Kruggles | timeline score: 2 | |
| Mar 12, 2019 at 21:26 | answer | added | George Ostrouchov | timeline score: 2 | |
| Mar 12, 2019 at 15:04 | answer | added | Sextus Empiricus | timeline score: 14 | |
| Mar 12, 2019 at 12:57 | comment | added | usεr11852 | In my case, I have found quantile regression much nicer to explain to non-technical people when the response variable is skewed (e.g. customer expenditure) and the introduction of a transformation/link-function step obscures the whole analysis. In that sense I would contest the assertion "median regression would give nearly identical results as linear regression" as being a bit oversimplifying; it does not, especially when dealing with potentially skewed response variables. | |
| Mar 12, 2019 at 10:16 | comment | added | Peter Flom | @MartijnWeterings Interesting questions. I don't know the answers. | |
| Mar 12, 2019 at 9:01 | comment | added | Sextus Empiricus | An answer could be like comparing the simple case of estimating a single population parameter, then showing that least squared errors performs better with Gaussian errors and least absolute residuals (using assumptions as well) performs better for different type of errors. But then, this question is about more complex linear models and the problem starts to be more complex and broad. The intuition of the simple problem (estimating a single mean/median) works for a bigger model, but by how much should it be worked out? And how to compare, robustness against outliers, distributions, computation? | |
| Mar 11, 2019 at 11:45 | comment | added | Christoph Hanck | As a further, but minor, point, one could maybe add the availbility of explicit, closed form solutions that are not available for, say, LAD, which may make such techniques less appealing for practitioners. | |
| S Mar 11, 2019 at 11:33 | history | suggested | Rafael Marazuela | CC BY-SA 4.0 |
I've added links to the Wikipedia.
|
| Mar 11, 2019 at 2:12 | review | Suggested edits | |||
| S Mar 11, 2019 at 11:33 | |||||
| Mar 10, 2019 at 16:55 | comment | added | whuber♦ | Given there are so many different variations of "linear regression," it would help to stipulate what you mean by this. Ditto for "median regression." In both cases, exactly how generally do you conceive these procedures? | |
| S Mar 10, 2019 at 16:48 | history | bounty started | Peter Flom | ||
| S Mar 10, 2019 at 16:48 | history | notice added | Peter Flom | Draw attention | |
| Mar 9, 2019 at 9:30 | comment | added | Christoph Hanck | I would argue that one important issue has been discussed in these two threads: stats.stackexchange.com/questions/153348/… and stats.stackexchange.com/questions/146077/… -- efficiency, and, possibly, even optimality under certain assumptions | |
| Mar 9, 2019 at 3:01 | history | tweeted | twitter.com/StackStats/status/1104215560995459077 | ||
| Mar 9, 2019 at 2:08 | comment | added | JustGettinStarted | To 'more familiar' i'd add 'interpretability' and 'stability', but for me one of the advantages of linear regression is what it tells you about the mean and how well that mean represents the sample population (residuals are very informative). Linear regression has as great value when its assumptions are met and good value when they are not met. | |
| Mar 8, 2019 at 16:01 | history | asked | Peter Flom | CC BY-SA 4.0 |