We propose a new methodology for filtering and forecasting the latent variance in a two-factor diffusion process with jumps from a continuous time perspective. For this purpose we use a continuous time Markov chain approximation with a finite state space. Essentially, we extend Markov chain...