DCSmooth 1.1.2
- Parallelization for DCS procedure added.
- Bandwidth selection of derivative estimation improved.
- Additional argument “trim” in surface.dcs for trimming the
boundaries of the plot.
- Data sets “returns.alv”, “returns.bmw”, “volumes.alv”, “volumes.bmw”
added.
- Bugs in estimation of derivatives (“KR” and “LP”) corrected and
corresponding error model are fixed.
- Default value in IPI_options$infl_par set to (2,1) for local
polynomial regression.
DCSmooth 1.1.0
- Estimation of variance model revised.
- set.options takes argument var_model = c(“iid”, “sarma_HR”,
“sarma_sep”, “sarma_RSS”, “sfarima_RSS”) instead of var_est. However,
var_est is still available for downstream compatibility, with the old
identifiers.
- The notation “QARMA” and “SARMA” has been unified to “SARMA” for all
spatial ARMA models.
- Functions qarma.est(), qarma.sim() have been replaced by sarma.est()
and sarma.sim(). They are still available and link to the new
functions.
- Added summary and print methods for classes “sarma” and
“sfarima”.
- Output “var_est” of dcs() contains now the complete estimation of
error terms of class “sarma” or “sfarima”.
- Arguments “model_order” and “order_max” controlling the orders of
the variance models are now passed to “set.options()” as additional
arguments in the functions ellipsis.
- Functions “kernel.assign()” and “kernel.list()” have been
added.
- New datasets “wind.nunn”, “wind.yuma”, “sun.nunn”, “sun.yuma”,
“returns.alv”, “returns.bmw”, “returns.sie”, “volumes.alv”,
“volumes.bmw”, “volumes.sie” added.
- Test for stationarity of SARMA models fixed.
DCSmooth 1.0.3
- New datasets “temp.nunn” and “temp.yuma” added.
- Estimation for error terms modeled after separable SARMA
included.
DCSmooth 1.0.2
- Functions for estimation and simulation of SFARIMA(p, q, d)
added.
DCSmooth 1.0.1
- Estimation for SFARIMA errors is now with ML estimation
DCSmooth 1.0.0
First release version * Estimation under SFARIMA errors (with
set.options(var_est = "lm")
) is only in experimental
state.