Timeline for Top $k$ List of Reasons to Close a Question Immediately
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 4 at 10:17 | history | edited | mkt | CC BY-SA 4.0 |
added 85 characters in body
|
| May 19, 2023 at 8:04 | comment | added | Nick Cox | Cultures differ. In dealing with some kinds of economic data, it seems routine to Winsorize heavily before the main analysis. Then there are follow-up questions (seen by me in other places) e.g. Should I Winsorize before transformation or after? Do I need to Winsorize all my variables or just those with outliers? I often challenge people to cite a textbook or authoritative review explaining why such Winsorizing is a good idea, and have yet to receive any reference by way of reply. Winsorizing to get a Winsorized mean or variance I get, but not this as a good route to a robust model.. | |
| Aug 24, 2022 at 14:04 | comment | added | gung - Reinstate Monica | Hmm, I hadn't noticed that it was a suggestion to answer here. I thought it was a suggestion to post the issue to Cross Validated Meta, like @Sycorax did here. | |
| Aug 24, 2022 at 11:29 | comment | added | mkt | @gung-ReinstateMonica I agree it's somewhat different from the rest. I posted this at whuber's suggestion: chat.stackexchange.com/transcript/message/61847310#61847310 . I could remove the dupe target suggestion if that's out of line. | |
| Aug 24, 2022 at 11:27 | comment | added | gung - Reinstate Monica | I'm not sure this topic belongs as an answer to this thread. I would delete this and ask a new question about the issue. Let people discuss what is / should be the canonical thread for the topic. | |
| Aug 24, 2022 at 9:17 | history | answered | mkt | CC BY-SA 4.0 |