Deriving Multi-Dimensional Weights
In the Couret and Venter paper, the derivation of the optimal credibility weights takes up ~2 pages, but there is no example of how to derive the optimal credibility weight. I have a difficult time understanding the covariance matrices presented, since there's no example of how I would do the calculations to get the optimal weight. The wiki summary in the "Credibility Considerations" section mentions that we should know the EPV and VHM formulas and how to apply them.
Do you think this is a likely or possible exam topic? If you think this is likely to be on the exam, would it be possible to get a small example of how to work through the calculations?
Comments
In our opinion this topic is quite unlikely but is possible in a very small example. We should probably revisit the wiki text to place less emphasis on memorizing the formulas for EPV and VHM as they're sufficiently complicated that hopefully the CAS would give them to you on the exam if needed.
Focus on knowing the structure of the matrix equation and where the individual components come from. We'll try to add an example in the future.