More on the project

Some more points that has been a bit unclear:

Exercise 2 d): In general the conditional distribution can have variances that

changes with sites, which is not assumed in model (*). Here you can either assume the variances are equal in the beginning or you can extend model (*) tol allow for different variances..

Exercise 3: Assume the epsilon-processes are independent in time

Exercise 3 j): Make some appropriate assumptions on alpha at the first time point.