correlation estimate from incomplete time series
Posted: August 2nd, 2006, 3:03 pm
I need to estimate a correlation between time series that have been marked at different times. Essentialy there is liquid data that is marked often, and some less liquid that is left the same until it is remarked.What is the optimal way to get a (valid) correlation matrix without dropping too many points ?NB: I don't give a toss about what your first thought about this problem is. I am asking for a known solution that has been tested and proved better than the naive one.