Functional Principal Component Analysis (FPCA)

Performs FPCA to estimate the density function for each subject.

PP_FPCA_new(
  t,
  h1 = NULL,
  h2 = NULL,
  N = NULL,
  bw = "ucv",
  Tend = 1,
  ngrid = 101,
  K.select = c("PropVar", "PPIC"),
  Kmax = 10,
  propvar = 0.85,
  density.method = c("kernel", "local linear"),
  polybinx = FALSE,
  derivatives = TRUE
)

Arguments

t	named vector; standardized longitudinal encounter times for a code with the corresponding patient name. They should be less than or equal to 1.
h1, h2	integers; bandwidth used to estimate the mean intensity function and the covariance function. Default=null.
N	named vector; the number of observed event with the corresponding patient name.
bw	a character string; bandwidth estimating method when h1 and h2 are null. Default="ucv", but can also be "nrd0", "nrd", "bcv","SJ-dpi" and "SJ-ste".
Tend	numeric; the upper bound of the encounter time for the estimated density function. Default=1.
K.select	characters indicating which method to choose the number of principal components K. Default is K.select="PropVar", and K.select="PPIC" is also available.
propvar	a proportion of variation used to select number of FPCs. Default is 0.85.
density.method	a character string; the method of estimating density function when K.select="PPIC". Default is "kernal", but can also be "local linear".
polybinx	logical; if use the same partition (x) for the polynomial regression when density.method="local linear". Default is FALSE.
derivatives	logical; whether to estimate the first derivatives of the density function. Default is TRUE.
n.grid	an integer value for grid points used in estimating covariance function g. Default is 101.