estimate.pred.dens: Estimate Predictive / Conditional Density

Description Usage Arguments

View source: R/er-functions.R

Description

estimate.pred.dens estimates the predictive / conditional density of order p associated with the scalar time series x using np.

Usage

1
2
estimate.pred.dens(x, p, kerntype = "gaussian", bwtype = "fixed",
  bwmethod = "cv.ml", bwuse = NA, nmulti = 1)

Arguments

x

A continuous-valued scalar time series.

p

The order of the conditional density.

kerntype

The kernel type, one of 'guassian', 'uniform', 'epanechnikov'.

bwtype

The bandwidth type, one of 'fixed', 'generalized_nn', 'adaptive_nn'.

bwmethod

The method used to choose the bandwidths, one of cv.ml, cv.ls, normal-reference. Defaults to cv.ml to perform LOOCV on the specific entropy rate.

bwuse

Specify a fixed bandwidth to use across all lags, or NA to estimate the bandwidths in a data-driven way.

nmulti

The number of random initial conditions to use by the optimizer for choosing the bandwidths.


ddarmon/spenra documentation built on July 11, 2020, 4:41 a.m.