cs_per: Generate design matrix for periodic cubic splines

In peRiodiCS: Functions for Generating Periodic Curves

Description

Generate design matrix for periodic cubic splines.

Usage

cs_per(x, knots = NULL, nk = 5, xmax = max(x, na.rm = TRUE),
  xmin = min(x, na.rm = TRUE))

Arguments

x	numerical x values to transform to new basis
knots	vector with locations of the knots of the spline
nk	number of knots, used only if the knots are not specified, overridden otherwise
xmax	value of the (theoretical) minimum of x
xmin	value of the (theoretical) maximum of x

Examples

# load example data; see help("viral_east_mediteranean")
data("viral_east_mediteranean")

# calculate location of knots to use
Knots <- 
 Hmisc::rcspline.eval(x = viral_east_mediteranean$EpiWeek,
                      nk = 5, knots.only = TRUE)

# model viral infections vs weeks
model <- glm(RSV ~ cs_per(EpiWeek, knots = Knots), data = viral_east_mediteranean)

# plot model (with many points, to make it smooth)
plot_per_mod(Model = model, XvarName = "EpiWeek", Smooth = TRUE)

peRiodiCS documentation built on May 2, 2019, 11:10 a.m.