Skip to main content
7 events
when toggle format what by license comment
Mar 30, 2021 at 15:23 history reopened whuber
Mar 30, 2021 at 1:25 history closed Stephan Kolassa
CommunityBot
Duplicate of Online estimation of variance with limited memory
Mar 29, 2021 at 12:43 comment added whuber One (amusing) solution method is to view this as a combination of a dataset of size $n$ and another one of size $-1,$ enabling you to apply (without change) formulas for updating moments when datasets are combined.
Mar 29, 2021 at 9:31 review Close votes
Mar 29, 2021 at 12:43
Mar 29, 2021 at 9:16 history edited Stephan Kolassa
edited tags
Mar 29, 2021 at 9:16 comment added Stephan Kolassa The question of updating the sample variance as one (or more) new datapoints becomes available is one of online-algorithms. So you can simply take an established algorithm and change it to yield the "old" variance based on the "new" variance, the data point to be removed and supplementary information like the "old" and "new" means.
Mar 29, 2021 at 8:06 history asked Mistakamikaze CC BY-SA 4.0