Normalizing constant of auto-regressive order 2 model

March 18, 2022 by Kinh Nguyen

Using the constant to correct for rank-deficiency interaction term when fitting model with TMB, such as a SEPARABLE effect between ICAR and auto-regressive order 2 (AR(2)). It can be extended for higher-order AR as well.

The AR(2) process here is stationary. The two correlation parameters are $\phi_1, \phi_2$, from this we can compute the AR2 graph and then the precision matrix $\mathbf Q$ \(x \sim AR(2)\) then \(f(x | \mu, \sigma, Q) = |Q| (Q^{-2})^{Q^{-2}} / (\sigma x \Gamma(Q^{-2})) \exp(Q^{-2}(Q w - \exp(Qw)))\) \(f(x | a, b, k) = (b / \Gamma(k)) (x^{bk -1} / a^{bk}) \exp(-(x/a)^b)\)