|
BGonline.org Forums
Conclusion
Posted By: rambiz In Response To: Conclusion (Maik Stiebler)
Date: Sunday, 14 August 2016, at 9:14 p.m.
Thank you very much for your clean and comprehensive derivation.
Allow me to pose a few questions:
1. Where does the "no bias" condition kicks in!?
2. Why optimize for a given correlation between X and Y? I think, cor_{X,Y} is rather a "design parameter" and not fixed.
3. For a given cor_{X,Y}, you showed that maximum gain in VR is achieved, when sig_Y/sig_X=cor_{X,Y}.
Can we still deviate from this condition?
4.What do you want to fine-tune except pushing down the variance!?
|
BGonline.org Forums is maintained by Stick with WebBBS 5.12.