Multifactor Vasicek model using Kalman Filter
Posted: November 28th, 2013, 10:48 am
I am trying to implement Vasicek model using Kalman filter using the research paper published by Bank of Canada http://www.bankofcanada.ca/en/res/wp/2001/wp01-15a.pdf I am facing problem in maximum likelihood function calculation my implementation is as follows measurement eqn :- Z = A + H *Y Variance Z = H *VarY*H' Error = Z(observed)- Z Kalman Gain = VarY * H' * Inv (VarZ) Y= Y+Kalman Gain*error VarY(tmp) = (I- Kalman gain*H) VarY Y = C + F Y Var Y = Var Y - F*VarY(tmp)'*F' + Qnow in likelihood function ( nlog(2pi)/2 - .5 ( in det(VarY) + error'* inv VarY*error)I am confused about the highlighted term whether usi VarY ( as mention in paper) is giving me matrix dimension error.using VAR Z is not giving me matrix error but I am not sure this is right approach or not.I am also attaching matlab code for your reference.I am really struggling with it.Please help