Calculates the optimal span for a loess spline

The optimal span is calculated based upon the bayesian information criterion (BIC).

optimal_span(
  y,
  x = NULL,
  weights = NULL,
  step = 0.01,
  label = NULL,
  plot = FALSE
)

Arguments

y

a vector with measurement values to smooth

x

a vector with dates / time steps

weights

optional values to weigh the loess fit with

step

span increment size

label

title to be used when plotting function output

plot

plot visual output of the optimization routine

Value

Returns an optimal span to smooth a provided vector using the `loess()` smoother.

Examples

if (FALSE) {
# Internal function only, should not be used stand-alone.
l <- sin(seq(1,10,0.01))
l <- l + runif(length(l))
optimal_span(l, plot = TRUE)
}