State Space Model Initial conditions

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

AltynayUssem
Posts: 1
Joined: Thu May 25, 2017 3:22 am

State Space Model Initial conditions

Postby AltynayUssem » Fri May 26, 2017 8:06 am

Hello!
I am very new to the State Space modelling, and probably will ask easy questions.
I have @signal exrate = c(1) + b1*(brent) + c(5)*baserate + [var=exp(c(2))]. and @state b1= b1(-1) + [var=exp(c(3))]
what is c(2)? I cannot understand
Also in setting initial values for example using param c(1) 0.0006 c(2) -8.5, c(3) -7 , should these 0.0006, -8.5 be the initial values of each time series for example for c(1) this is the first value of exrate? or should I take as initial values the means of the time series? and a special note on c(2) which inital value should I put for it as it is specified as [var=exp(c(2))]?
Sorry and
Thanks,
Altyn

startz
Non-normality and collinearity are NOT problems!
Posts: 3382
Joined: Wed Sep 17, 2008 2:25 pm

Re: State Space Model Initial conditions

Postby startz » Fri May 26, 2017 8:40 am

AltynayUssem wrote:Hello!
I am very new to the State Space modelling, and probably will ask easy questions.
I have @signal exrate = c(1) + b1*(brent) + c(5)*baserate + [var=exp(c(2))]. and @state b1= b1(-1) + [var=exp(c(3))]
what is c(2)? I cannot understand
Also in setting initial values for example using param c(1) 0.0006 c(2) -8.5, c(3) -7 , should these 0.0006, -8.5 be the initial values of each time series for example for c(1) this is the first value of exrate? or should I take as initial values the means of the time series? and a special note on c(2) which inital value should I put for it as it is specified as [var=exp(c(2))]?
Sorry and
Thanks,
Altyn

The c vector gives initial guesses of parameter values--nothing to do with initial values of a series. c(2) is a guess as to the log of the variance of shocks to exrate. c(1) is a guess as to the intercept in the equation.


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 9 guests