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 |
library("plot") library("kalman") c.f. original kemitor by P.Franek for more examples 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")
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)]
library("plot") library("kalman") c.f. original kemitor by P.Franek for more examples 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")
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)]