Using a principal component constructed from the sample space, we simulate
null distributions with univariate Normal distributions using pcn_simulate
.
Then a subset of these distributions is chosen using pcn_select
.
pcn_simulate(data, n.sim = 50)
pcn_select(data.sim, cl, type = c("rep", "range"), int = 5)
data matrix with rows as samples, columns as features
The number of simulated datasets to simulate
an object from pcn_simulate
vector of cluster memberships
select either the representative dataset ("rep") or a range of datasets ("range")
every int
data sets from median-ranked data.sim
are taken.
Defaults to 5.
pcn_simulate
returns a list of length n.sim
. Each element is a
simulated matrix using this "Principal Component Normal" (pcn) procedure.
pcn_select
returns a list with elements
ranks
: When type = "range"
, ranks of each extracted dataset shown
ind
: index of representative simulation
dat
: simulation data representation of all in pcNormal