Python and Jupiter

Using the data in Ex.3 of the case, regress HFRI on each index using a window of 10 years to forecast the 12 monthly returns for the following year. So, you
can use the data from 1996 to 2005 to forecast the return of the 12 months of 2006. Repeat the same exercise year by year until 2011. For the whole period
as well as 2008-9 and 2009-11:
A.What is the average monthly return, volatility, Sharpe ratio, skewness and kurtosis of HFRI and your forecasts?
B.How well do predicted returns correlate with HFRI?
C.Repeat the same exercise with a 3 years and 5 years window. Is there any improvement?
D.How much fees could HFRI replication charge before falling below S&P Sharpe (you can answer this question for the whole period)?