Displaying 2 results from an estimated 2 matches for "degruyter".
2017 Oct 17
4
uniform sampling without replacement algorithm
...where the sample is selected from
1...n, since the algorithm has to create a vector of length n. So when
the sample size is much lesser than n, the algorithm is not effective.
Algorithms with average complexity O(s log s), were s is the sample
size, were described long ago. E.g. see
https://www.degruyter.com/view/j/mcma.1999.5.issue-1/mcma.1999.5.1.39/mcma.1999.5.1.39.xml
Here the Tree algorithm has complexity O(s log s). I suppose that there
may be algorithms with complexity close to s. Is somebody planning to
implement some more effective algorithm?
2017 Oct 18
0
uniform sampling without replacement algorithm
...elected from 1...n,
> since the algorithm has to create a vector of length n. So when the sample
> size is much lesser than n, the algorithm is not effective. Algorithms with
> average complexity O(s log s), were s is the sample size, were described
> long ago. E.g. see
> https://www.degruyter.com/view/j/mcma.1999.5.issue-1/mcma.
> 1999.5.1.39/mcma.1999.5.1.39.xml
> Here the Tree algorithm has complexity O(s log s). I suppose that there
> may be algorithms with complexity close to s. Is somebody planning to
> implement some more effective algorithm?
>
> ________________...