Add a strict sliding‑window estimate to the eda_dens function

[mgimond]

For pedagogical reasons, I will add a sliding-window estimator (for a rectangular kernel) to eda_dens. The estimate will take on the form:

strict_kde <- function(x, y, bw) {
  n <- length(y)
  sapply(x, function(x0) {
    sum(abs(y - x0) <= (bw/2) ) / (n * bw)
  })
}

where x is a vector of x values for which density values are to be computed , y is the batch of values whose density is estimated and bw is the full bandwidth of the rectangular kernel.