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

Quantlet: 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, 21.9.2000