Grocer Function
NAME
filter_like - Kalman filter log-likelihood
CALLING SEQUENCE
[llik]=filter_like(param,func,y,x,F,z)
PARAMETERS
Input
-
param = a vector of parameters (sqrt of variances)
-
func = the function which transforms the parameters into the matrix of variances (Q and R)
-
y = (nx1) data vector
-
x = (nxk) data matrix
-
F = transition matrix
-
z = (nxl) data matrix
Output
-
llik = - 2*log-likehood (+n*log(2*%pi))
DESCRIPTION
Generates model loglikelihood in a kalman model:
y(t) = X(t)*B(t) + Z(t)*A + e(t), e(t) = N(0,R)
B(t) = F*B(t-1) + v(t), v(t) = N(0,Q)
EXAMPLE
Filter_like is used by the function kalman:
oresult = maxlik(filter_like,grocer_param,grocer_func,grocer_y,grocer_x,grocer_F, varargin(:))
AUTHOR
Eric Dubois 2002