- could you pls elaborate the comment you added to Case D of your last spreadsheet:
Case D appears to show that there is only two degrees of freedom because there are only three maximum errors that are equal. With 5 variables there would be 6 of them.
- you tried to tackle the problem via a variance-covariance matrix...would you possibly have any further insights into this approach (I do not necessarily need to use LSq; it was the one approach I was reasonably familiar with but I am happy to try out other optimization approaches)This seems appropriate for the "noisy data" case where the matrix is the noisy data. Wikipedia treats this case pretty well at:
- Regarding the idea of optimizing the (node) weights): I am not per se against this (apologies if it came across as such), I just have problems understanding the practical implications of adjusting the node weights...if I can translate this into an improved hedge portfolio I am fully for it. Would you have an example of using the variance-rules to guide the specification of the weights?See: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CB4QFjAA&url=http%3A%2F%2Fstat.psu.edu%2F~rho%2Fstat462%2FApril16.doc&ei=U9ZXVL6iMIb1iQKNloC4Ag&usg=AFQjCNFdlyVByUy-jTL812mjZciCKLXYQQ&bvm=bv.78677474,d.cGE
- In the spreadsheet attached to your poste on the 28Oct (LeastSquares-Matrices) you calculate the least squares solution directly...would there be any benefits/advantages over the solver-generated solution?Perhaps. I rather like the Solver and I'd guess it's a lot easier to set up. But, if compute time is critical then you could try both and see.
If you are experiencing a similar issue, please ask a related question
Join the community of 500,000 technology professionals and ask your questions.
Connect with top rated Experts
20 Experts available now in Live!