Updating the sample variance, sample by sample

Asked Oct 25 '12 at 16:45

Active Oct 25 '12 at 16:49

Viewed 44 times

Possible Duplicate:
Computation of new standard deviation using old standard deviation after change in dataset

I know the sample variance of n samples and its mean and n itself. Now I would like to add one more sample so I have n + 1. Is there a way how to add the sample without recalculation the whole term. (computational efficient). Approximation would be fine too.

edited Apr 13 '17 at 12:44

Community

asked Oct 25 '12 at 16:45

Davoud Taghawi-Nejad

Do you also know the size of each sample? – Peter Flom Oct 25 '12 at 16:49
1

I believe you're looking for an online algorithm for calculating variance. See [this link on wikipedia](http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm). – assumednormal Oct 25 '12 at 16:50
I assumed that "sample" referred to a single observation. If that's incorrect, disregard my previous comment. – assumednormal Oct 25 '12 at 16:51
a given sample at the beginning and than one sample each time. – Davoud Taghawi-Nejad Oct 25 '12 at 16:57
Max, you did answer my question. You could answer the usual way and I could accept it. (otherwise how to close this question?) – Davoud Taghawi-Nejad Oct 25 '12 at 17:00
1

For more information about such "online" approaches to updating statistical summaries, look at the threads tagged with [tag:online]. – whuber Oct 25 '12 at 17:57
Its not a duplicate the proposed link is about replaceing an observation, while this is about adding an observation. – Davoud Taghawi-Nejad Oct 26 '12 at 21:10

Updating the sample variance, sample by sample

0 Answers0