compute_scale_par: Compute scale parameter for error in kdde
In gbasulto/kerdec: Kernel deconvolution methods

Description Usage Arguments Details Value Author(s) Examples

This function computes the maximum likelihood estimator of the standard deviation for the error distribution provided as long as this error is normal or Laplace and the sample is either a sample or errors or a sample of differences of errors (which arises in panel data).

1	compute_scale_par(error_dist, error_smp, k, error_scale_par = NULL)

`error_dist`	An integer specifying the error distribution: 1 if non-parametric, 2 if Laplace, 3 if normal.
`error_smp`	Vector of sample of errors or differences of errors.
`k`	An integer specifying what is being provided as sample of errors: '1' for pure sample of errors and otherwise for differences of errors.
`error_scale_par`	NULL by default. It is is given, the function does nothing but return this value.

This function also computes the standard deviation when the error is approximated with the empirical characteristic function, since it is useful for bandwidth selection.

The (positive) standard deviation

Guillermo Basulto-Elias

## Difference of Laplace errors is provided:
sigma <- 5
smp <- rlaplace(n = 100, sd = sigma) - rlaplace(n = 100, sd = sigma)
kerdec:::compute_scale_par(2, smp, 2)