[Yuima-commits] r500 - in pkg/yuima: . man src
noreply at r-forge.r-project.org
noreply at r-forge.r-project.org
Sat Oct 29 23:03:26 CEST 2016
Author: lorenzo
Date: 2016-10-29 23:03:26 +0200 (Sat, 29 Oct 2016)
New Revision: 500
Modified:
pkg/yuima/DESCRIPTION
pkg/yuima/man/qmle.Rd
pkg/yuima/src/carmafilter.c
Log:
Added qmle doc,
simulate PPR
Modified: pkg/yuima/DESCRIPTION
===================================================================
--- pkg/yuima/DESCRIPTION 2016-10-29 14:41:35 UTC (rev 499)
+++ pkg/yuima/DESCRIPTION 2016-10-29 21:03:26 UTC (rev 500)
@@ -1,7 +1,7 @@
Package: yuima
Type: Package
Title: The YUIMA Project Package for SDEs
-Version: 1.2.9
+Version: 1.3.0
Depends: R(>= 2.10.0), methods, zoo, stats4, utils, expm, cubature, mvtnorm
Imports: Rcpp (>= 0.12.1)
Author: YUIMA Project Team
Modified: pkg/yuima/man/qmle.Rd
===================================================================
--- pkg/yuima/man/qmle.Rd 2016-10-29 14:41:35 UTC (rev 499)
+++ pkg/yuima/man/qmle.Rd 2016-10-29 21:03:26 UTC (rev 500)
@@ -46,7 +46,7 @@
\item{start}{initial values to be passed to the optimizer.}
\item{fixed}{for conditional (quasi)maximum likelihood estimation.}
\item{joint}{perform joint estimation or two stage estimation? by default \code{joint=FALSE}.}
- \item{Est.Incr}{If the yuima model is an object of \code{\link{yuima.carma-class}} the \code{qmle} returns an object of \code{\link{yuima.carma.qmle-class}} or object of class \code{mle-class}. By default \code{Est.Incr="Carma.IncPar"}.}
+ \item{Est.Incr}{If the yuima model is an object of \code{\link{yuima.carma-class}} or \code{\link{yuima.cogarch-class}} the \code{qmle} returns an object of \code{\link{yuima.carma.qmle-class}}, \code{\link{cogarch.est.incr-class}},\code{\link{cogarch.est-class}} or object of class \code{mle-class}. By default \code{Est.Incr="NoIncr"}, alternative values are \code{IncrPar} and \code{Incr}. }
\item{aggregation}{If \code{aggregation=TRUE}, before the estimation of the levy parameters we aggregate the increments.}
\item{threshold}{If the model has Compund Poisson type jumps, the threshold is
used to perform thresholding of the increments.}
@@ -62,7 +62,6 @@
\code{lse} calculates least squares estimators of the drift parameters. This is
useful for initial guess of \code{qmle} estimation.
-
\code{quasilogl} returns the value of the quasi loglikelihood for a given
\code{yuima} object and list of parameters \code{coef}.
@@ -93,6 +92,10 @@
## COGARCH
Iacus S. M., Mercuri L. and Rroji E.(2015) Discrete time approximation of a COGARCH (p, q) model and its estimation. \href{http://arxiv.org/abs/1511.00253}{http://arxiv.org/abs/1511.00253}
+
+## CARMA
+
+Iacus S. M., Mercuri L. (2015) Implementation of Levy CARMA model in Yuima package. Comp. Stat. (30) 1111-1141. \href{http://link.springer.com/article/10.1007/s00180-015-0569-7}{http://link.springer.com/article/10.1007/s00180-015-0569-7}
}
\author{The YUIMA Project Team}
@@ -102,6 +105,7 @@
The function qmle uses the function \code{\link{CarmaNoise}} internally for estimation of underlying Levy if the model is an object of \code{\link{yuima.carma-class}}.
}
+
\examples{
#dXt^e = -theta2 * Xt^e * dt + theta1 * dWt
diff.matrix <- matrix(c("theta1"), 1, 1)
@@ -168,7 +172,6 @@
\dontrun{
-
###multidimension case
##dXt^e = - drift.matrix * Xt^e * dt + diff.matrix * dWt
diff.matrix <- matrix(c("theta1.1","theta1.2", "1", "1"), 2, 2)
@@ -224,7 +227,6 @@
opt3 at coef
summary(opt3)
-
quasilogl(yuima, param=list(theta2.1=0.8, theta2.2=0.2, theta1.1=0.7, theta1.2=0.1))
##system.time(
@@ -235,7 +237,6 @@
# carma(p=2,q=0) driven by a brownian motion without location parameter
-
mod0<-setCarma(p=2,
q=0,
scale.par="sigma")
@@ -259,10 +260,6 @@
summary(carmaopt0)
-
-
-
-
# carma(p=2,q=1) driven by a brownian motion without location parameter
mod1<-setCarma(p=2,
@@ -286,8 +283,6 @@
summary(carmaopt1)
-plot(carmaopt1)
-
# carma(p=2,q=1) driven by a compound poisson process where the jump size is normally distributed.
mod2<-setCarma(p=2,
@@ -313,8 +308,6 @@
summary(carmaopt2)
-plot(carmaopt2)
-
# carma(p=2,q=1) driven by a normal inverse gaussian process
mod3<-setCarma(p=2,q=1,
measure=list(df=list("rNIG(z, alpha, beta, delta1, mu)")),
@@ -343,8 +336,6 @@
summary(carmaopt3)
-plot(carmaopt3)
-
# Simulation and Estimation of COGARCH(1,1) with CP driven noise
# Model parameters
Modified: pkg/yuima/src/carmafilter.c
===================================================================
--- pkg/yuima/src/carmafilter.c 2016-10-29 14:41:35 UTC (rev 499)
+++ pkg/yuima/src/carmafilter.c 2016-10-29 21:03:26 UTC (rev 500)
@@ -266,7 +266,7 @@
/*if(rResult[1]>0){*/
rResult[0] += -0.5 * (log(rResult[1])+ Uobs * Uobs /rResult[1]);
/*}else{
- rResult[0] += -1000000;
+ rResult[0] += -10000000000;
}*/
/*printf("\n res %.5f", rResult[0]);*/
/* manual debug */
More information about the Yuima-commits
mailing list