Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: kalman
See also: kfilter ksmoother kem kemitor epscontnorm

Macro: kemitor2
Description: Simulates observations and states of a given state-space-model - just as kemitor by Petr Franek (quantlib times) - but this time also the states are returned. The state-space model is assumed to be in the following form:

y_t = H x_t + ErrY_t

x_t = F x_t-1 + ErrX_t

x_0 = mu


Usage: {X,Y} = kemitor2(T,mu,H,F,ErrY,ErrX)
Input:
T number of observations to be generated
mu n x 1 vector (starting point of the model)
H m x n matrix
F n x n matrix
ErrY T x m matrix of errors
ErrX T x n matrix of errors
Output:
X T x n matrix of generated states, T is the number of generated states, n is the dimension of generated states
Y T x m matrix of generated time series, T is the number of generated observations, m is the dimension of generated time series

Example:
library("plot")
library("kalman")
randomize(0)
T = 100
mu=0
H = 1
F = 1
mid=0
Cid=1
mcont=10
Ccont=0.1
eps=0.1
ErrX = normal(T)
ErrY = epscontnorm(T,eps,mid,Cid,mcont,Ccont,0)
sim = kemitor2(T,mu,H,F,ErrY,ErrX)
state = (1:100)~(sim.X)
obs= (1:100)~(sim.Y)
ind = (ErrY.Ind)
ind=ind/ind*0 ;           sets all "0" to missing and all "1" to 0 
ind = (1:100)~ind
setmaskp(ind,4, 3, 8)  ;       on all contam. instances a red flag is plotted
state=setmask(state,"line","blue","thin")          
obs=setmask(obs,"line","green","thin")          
disp = createdisplay(1,1)
show(disp,1,1,state,obs,ind)
setgopt(disp,1,1, "title", "State Space Model (1 - dim steady state) under AO")
Result:
Generates a simulation of length 100 of a steady state model (i.e. H=R=Q=F=1)
under a convex contaminated ErrY (AO) [radius=0.1, K=N(10,0.1)]
Example:
library("plot")
library("kalman")
randomize(0)
T = 100
mu=0
H = 1
F = 1
mid=0
Cid=1
mcont=10
Ccont=0.1
eps=0.1
ErrY = normal(T)
ErrX = epscontnorm(T,eps,mid,Cid,mcont,Ccont,0)
sim = kemitor2(T,mu,H,F,ErrY,ErrX)
state = (1:100)~(sim.X)
obs= (1:100)~(sim.Y)
ind = (ErrX.Ind)
ind=ind/ind*0 ;           sets all "0" to missing and all "1" to 0 
ind=(1:100)~ind
setmaskp(ind,4, 3, 8)  ;       on all contam. instances a red flag is plotted
state=setmask(state,"line","blue","thin")          
obs=setmask(obs,"line","green","thin")          
disp = createdisplay(1,1)
show(disp,1,1,state,obs,ind)
setgopt(disp,1,1, "title", "State Space Model (1 - dim steady state) under IO")
Result:
Generates a simulation of length 100 of a steady state model (i.e. H=R=Q=F=1)
under a convex contaminated ErrX (IO) [radius=0.1, K=N(10,0.1)]

Library: kalman
See also: kfilter ksmoother kem kemitor epscontnorm

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Author: P.Ruckdeschel 991018
(C) MD*TECH Method and Data Technologies, 17.8.2000