Skip to main content
13 events
when toggle format what by license comment
S Jul 13, 2021 at 7:30 history bounty ended Trevor Hansen
S Jul 13, 2021 at 7:30 history notice removed Trevor Hansen
Jul 13, 2021 at 7:17 vote accept Trevor Hansen
Jul 12, 2021 at 19:54 answer added Ben timeline score: 2
Jul 12, 2021 at 1:42 answer added Subrata Pal timeline score: 1
S Jul 9, 2021 at 10:58 history bounty started Trevor Hansen
S Jul 9, 2021 at 10:58 history notice added Trevor Hansen Draw attention
Jul 8, 2021 at 0:00 history tweeted twitter.com/StackStats/status/1412924404573655042
Jul 7, 2021 at 18:22 comment added Trevor Hansen Thanks @whuber. Will probably need to consider the distributional family and the problem at hand indeed. I am working on a problem with few observations (~30) and expect that the higher order moments will be a less stable approach. Will assess what works best for our problem numerically, was just hoping that there were some general materials/papers/analyses to help me on my way.
Jul 7, 2021 at 15:59 comment added whuber I think your observation "easier to solve analytically" may be the best explanation. Indeed, why restrict yourself to integral moments? An investigation of that possibility indicates one might need to analyze any given distributional family to determine what an optimal combination of moments would be. This suggests there is no universal answer to the question of finding the most accurate (or lowest-loss) choice of moments.
Jul 7, 2021 at 13:59 comment added Trevor Hansen @ChristophHanck the post you link to compares central second moment with raw second moment (these both have the same order). I do not see how that helps the question of absolute first central moment and central second order moment, or am I missing something?
Jul 7, 2021 at 13:38 comment added Christoph Hanck related: stats.stackexchange.com/questions/209328/…
Jul 7, 2021 at 10:37 history asked Trevor Hansen CC BY-SA 4.0