Displaying 2 results from an estimated 2 matches for "sucbin".
Did you mean:
subin
2002 Aug 08
0
RE: rmultinom
...er box).
However, the successive binomial version is slower when observations are
sparse, presumably because it spends most of its time filling in zeros that
tabulate() doesn't worry about.
test> # Low dimension, many observations per box
test> length( ranvec)
[1] 7
test> system.time( sucbin.rmultinom( 1000, 1000, ranvec))
[1] 0.02 0.00 0.02 NA NA
test> system.time( tab.rmultinom( 1000, 1000, ranvec))
[1] 0.19 0.00 0.18 NA NA
test> # High dimension, few observations per box
test> length( ranvec.long)
[1] 1000
test> system.time( sucbin.rmultinom( 10000, 10, ranvec....
2002 Aug 08
0
RE: rmultinom
...sive binomial version is slower when observations are
>sparse, presumably because it spends most of its time filling in zeros that
>tabulate() doesn't worry about.
>
>test> # Low dimension, many observations per box
>test> length( ranvec)
>[1] 7
>test> system.time( sucbin.rmultinom( 1000, 1000, ranvec))
>[1] 0.02 0.00 0.02 NA NA
>test> system.time( tab.rmultinom( 1000, 1000, ranvec))
>[1] 0.19 0.00 0.18 NA NA
>
>test> # High dimension, few observations per box
>test> length( ranvec.long)
>[1] 1000
>test> system.time( sucbi...