MixARgen-class {mixAR} | R Documentation |
Class "MixARgen"
Description
A class for MixAR models with arbitrary noise distributions. "MixARgen"
inherits from "MixAR"
.
Objects from the Class
Objects can be created by calls of the form new("MixARgen",
dist, ...)
or mixARgen(...)
. The two forms are completely
equivalent. The latter is available from version 0.19-15 of package
MixAR.
Slots
Most slots are inherited from class "MixAR"
.
prob
:the mixing probabilities,
"numeric"
.order
:the AR orders,
"numeric"
.shift
:intercept terms,
"numeric"
.scale
:scaling factor,
"numeric"
.arcoef
:-
autoregressive coefficients, an object from class
"raggedCoef"
containing one row for each mixture component. dist
:Object of class
"list"
, representing the noise distributions. The list contains one element for each component of the MixAR model or a single element if the noise distribution is the same for all components.If the distributions do not contain parameters (e.g. Gaussian or
t_4
) it is sufficient to give the list of functions in the elementdist
of the list.If the distributions do contain parameters the recommended arrangement is to give a list with components
generator
andparam
, such that a callgenerator(param)
should produce the required list of distributions.This is not finalised but if changed, backward compatibility with existing objects will be maintained.
Extends
Class "MixAR"
, directly.
Methods
- get_edist
signature(model = "MixARgen")
: ...- initialize
signature(.Object = "MixARgen")
: ...- lik_params
signature(model = "MixARgen")
: ...- mix_cdf
signature(model = "MixARgen", x = "missing", index = "missing", xcond = "numeric")
: ...- mix_cdf
signature(model = "MixARgen", x = "numeric", index = "missing", xcond = "numeric")
: ...- mix_cdf
signature(model = "MixARgen", x = "numeric", index = "numeric", xcond = "missing")
: ...- mix_pdf
signature(model = "MixARgen", x = "missing", index = "missing", xcond = "numeric")
: ...- mix_pdf
signature(model = "MixARgen", x = "numeric", index = "missing", xcond = "numeric")
: ...- mix_pdf
signature(model = "MixARgen", x = "numeric", index = "numeric", xcond = "missing")
: ...- noise_dist
signature(model = "MixARgen")
: ...- noise_params
signature(model = "MixARgen")
: ...- noise_rand
signature(model = "MixARgen")
: ...
Examples
showClass("MixARgen")
exampleModels$WL_ibm_gen@dist
noise_dist(exampleModels$WL_ibm_gen, "cdf")
noise_dist(exampleModels$WL_ibm_gen, "pdf")
noise_dist(exampleModels$WL_ibm_gen, "pdf", expand = TRUE)
noise_dist(exampleModels$WL_ibm_gen, "cdf", expand = TRUE)
## data(ibmclose, package = "fma") # for `ibmclose'
pdf1 <- mix_pdf(exampleModels$WL_ibm, xcond = as.numeric(fma::ibmclose))
cdf1 <- mix_cdf(exampleModels$WL_ibm, xcond = as.numeric(fma::ibmclose))
gbutils::plotpdf(pdf1, cdf = cdf1, lq = 0.001, uq = 0.999)
pdf1gen <- mix_pdf(exampleModels$WL_ibm_gen, xcond = as.numeric(fma::ibmclose))
cdf1gen <- mix_cdf(exampleModels$WL_ibm_gen, xcond = as.numeric(fma::ibmclose))
gbutils::plotpdf(pdf1gen, cdf = cdf1gen, lq = 0.001, uq = 0.999)
length(fma::ibmclose)
cdf1gena <- mix_cdf(exampleModels$WL_ibm_gen, xcond = as.numeric(fma::ibmclose)[-(369:369)])
pdf1gena <- mix_pdf(exampleModels$WL_ibm_gen, xcond = as.numeric(fma::ibmclose)[-(369:369)])
gbutils::plotpdf(pdf1gena, cdf = cdf1gena, lq = 0.001, uq = 0.999)
pdf1a <- mix_pdf(exampleModels$WL_ibm, xcond = as.numeric(fma::ibmclose)[-(369:369)])
cdf1a <- mix_cdf(exampleModels$WL_ibm, xcond = as.numeric(fma::ibmclose)[-(369:369)])
gbutils::plotpdf(pdf1a, cdf = cdf1a, lq = 0.001, uq = 0.999)
cdf1gena <- mix_cdf(exampleModels$WL_ibm_gen, xcond = as.numeric(fma::ibmclose)[-(369:369)])
cond_loglik(exampleModels$WL_ibm, as.numeric(fma::ibmclose))
cond_loglik(exampleModels$WL_ibm_gen, as.numeric(fma::ibmclose))
ts1gen <- mixAR_sim(exampleModels$WL_ibm_gen, n = 30, init = c(346, 352, 357), nskip = 0)
plot(ts1gen)
plot(mixAR_sim(exampleModels$WL_ibm_gen, n = 100, init = c(346, 352, 357), nskip = 0),
type = "l")
plot(diff(mixAR_sim(exampleModels$WL_ibm_gen, n = 100, init = c(346, 352, 357), nskip = 0)),
type = "l")
noise_dist(exampleModels$WL_ibm_gen, "Fscore")
prob <- exampleModels$WL_ibm@prob
scale <- exampleModels$WL_ibm@scale
arcoef <- exampleModels$WL_ibm@arcoef@a
mo_WLt3 <- new("MixARgen", prob = prob, scale = scale, arcoef = arcoef,
dist = list(fdist_stdt(3)))
mo_WLt30 <- new("MixARgen", prob = prob, scale = scale, arcoef = arcoef,
dist = list(fdist_stdt(30)))