|
Hello everyone,
I am doing a project which involves stock returns data set. As far as I know financial time series have serial correlation and volatility clustering. I have to compute the sample covariance matrix for different historical windows, which is then used in a portfolio optimisation problem, i.e I use the historical covariance matrix to forecast volatility. If I use the "cov" function I will probably end up with biased matrices and a lot of noise. The literature on this topic suggests Random Matrix Theory, Principal Component Analysis and quite a few other filtering methods. My question is, which one should I use to fix my covariance matrices and how do I do that in Matlab (functions, code)? I found something in the user guide about "wishrnd" but i am not sure how to use it and if it is any good. The idea is that I need to be able to find a filtered matix from the "noisy" one, that can be used
in the optimisation problem. Thanks a lot for your help in advance.
|
Feed for this Thread