Consider the problem of approximating a continuous analog signal
with a discrete digital signal . To do this, we may divide the
range of into discrete intervals, , , etc.
Then, if takes on a value in the interval , we assign
the value to the discrete signal (predictor) , and so on for
the other intervals.
Suppose that is distributed on according to the PDF
. If we discretize by picking ,
and , how should we pick and
to generate the best predictor of (that is, to minimize
?
In this case, what is the expectation?
You can earn partial credit on this problem.