Sample a fraction of shapes from a Momocs object. See examples and ?dplyr::sample_n.
Arguments
- tbl
a Momocs object (Coo, Coe)
- size
numeric (0 < numeric <= 1) the fraction of shapes to select
- replace
logical whether sample should be done with ot without replacement
- fac
a column name if a
$facis defined; size is then applied within levels of this factor- ...
additional arguments to dplyr::sample_frac and to maintain generic compatibility
Note
the resulting fraction is rounded with ceiling.
See also
Other handling functions:
arrange(),
at_least(),
chop(),
combine(),
dissolve(),
fac_dispatcher(),
filter(),
mutate(),
rename(),
rescale(),
rm_harm(),
rm_missing(),
rm_uncomplete(),
rw_fac(),
sample_n(),
select(),
slice(),
subsetize()
Examples
# samples 50% of the bottles no matter their type
sample_frac(bot, 0.5)
#> Out (outlines)
#> - 20 outlines, 159 +/- 23 coords (in $coo)
#> - 2 classifiers (in $fac):
#> # A tibble: 20 × 2
#> type fake
#> <fct> <fct>
#> 1 beer d
#> 2 beer d
#> 3 beer d
#> 4 whisky a
#> 5 beer c
#> 6 beer d
#> # ℹ 14 more rows
#> - also: $ldk
# 80% bottles of beer and of whisky
table(sample_frac(bot, 0.8, fac="type")$fac)
#> fake
#> type a b c d
#> beer 0 0 8 8
#> whisky 8 8 0 0
# bootstrap the same number of bootles of each type but with replacement
table(names(sample_frac(bot, 1, replace=TRUE)))
#>
#> amrut ballantines bushmills caney chimay corona
#> 1 1 2 1 2 1
#> dalmore duvel famousgrouse franziskaner glendronach grimbergen
#> 2 1 1 4 1 1
#> guiness highlandpark jb jupiler kingfisher latrappe
#> 2 2 1 1 1 2
#> magallan oban oldpotrero pecheresse sierranevada westmalle
#> 4 1 2 2 1 1
#> wildturkey yoichi
#> 1 1