Kalman Filter
Posted: Wed Apr 17, 2013 5:24 am
Hello all,
I'm trying to estimate a Kalman Filter but I'm having some trouble setting up the right state-space model.
The point of the estimation is to estimate the price of 6 stocks, using two factors. One factor is a common factor (so its the same for all stocks) and the other is an individual one.
so the measurement equation should be: (i is the stock and t the time period)
p(t,i) = lambda1*FC(t) + lambda2*FI(t,i) + error(t,i)
the two factors should be defined in the space equation as AR(1) functions so
FC(t)= a1*FC(t-1), FI(t)=a2*FI(t-1)
where a is the autocorrelation parameter
can someone help me with this please?
I'm trying to estimate a Kalman Filter but I'm having some trouble setting up the right state-space model.
The point of the estimation is to estimate the price of 6 stocks, using two factors. One factor is a common factor (so its the same for all stocks) and the other is an individual one.
so the measurement equation should be: (i is the stock and t the time period)
p(t,i) = lambda1*FC(t) + lambda2*FI(t,i) + error(t,i)
the two factors should be defined in the space equation as AR(1) functions so
FC(t)= a1*FC(t-1), FI(t)=a2*FI(t-1)
where a is the autocorrelation parameter
can someone help me with this please?