John Dodson:hi there
  John Dodson:how are you getting on with the homework?
  John Dodson:do we need to use the conference line? If not, I will drop it.
  John Dodson:I do not hear anyone on the conference line, so I am going to
drop it and go over to Connect audio.
  sangya:yes
  Mo:yes
  John Dodson:I will save and post tonight's chat session for those who cannot
attend.
  John Dodson:Have you have a chance to start in on the Engle-Ng paper for next
week?
  sangya:I haven't yet. Have the grades been posted to moodle? i saw you added
the link
  sangya:but when i sign into moodle i don't see this class
  John Dodson:The module exists on Moodle2, it is FM5031_001F11, but I don't
see any grades yet.
  sangya:ok
  John Dodson:Sangya, I don't know what Moodle looks like from the student
perspective. Are you using Moodle2?
  sangya:Yes - when I follow your link, I get a message saying this moodle site
is unavailable to students. it says the instructor has to unhide it first.
  John Dodson:oh, ok. I will work on that
  sangya:k, thanks
  John Dodson:OK, I have changed the "availability" on the Moddle page. Does it
appear now, Sangya?
  Mo:John if i understood well the hmk, it has 2 sides which are one evaluation
on daily return and the other on daily vol, right?
  John Dodson:Yes, Mo. You can assume daily returns are normal and use the MLE
and C-R results for estimate and standard error for the two parameters.
  John Dodson:Technically, the term "volatility" refers to the geometric
Brownian motion model, which involves a normal but has a slightly different
interpretation for the parameters
  sangya:It appears on moodle now, thanks
  John Dodson:OK, thanks Sangya. I am new to Moodle.
  sangya:no problem
  sangya:could we just use MLE in matlab?
  John Dodson:probably, Sangya; but I am not familiar with the details of the
statistics package.
  sangya:ok. can you explain the se equation on slide 12 again? particularly
the diag part?
  John Dodson:The 'diag()' probably causes more confusion than it is worth. The
point is that, which most MLE's are joint estimators, sometimes we only care
about their marginal performances.
  John Dodson:The standard error (se) is the vector of the (marginal) standard
deviation of the estimator components
  John Dodson:To get standard deviation from covariance (matrix), we have to
first extract its diagonal.
  John Dodson:We can't just take the square root of the covariance, becaues
technically this would be something like the Cholesky matrix. But we only care
about the margins.
  sangya:that helps, thanks
  John Dodson:so the first 'diag()' pulls out the diagonal of the cov, the
second one turns it into a diagonal matrix with zero off-diagonal entries, and
the third one after the 'sqrt()' extracts the diagonal of the result.
  sangya:i had to try it in matlab step by step but now I get it :)
  Greg:John, have you had a chance to post some old final exams?
  John Dodson:no, not yet Greg. I will try to get some sample questions out
before my next visit on the 19th.
  Greg:Thanks, I appreciate it.
  John Dodson:Any final questions before we call it a night?
  Greg:Nope.  Have a good night.
  Xiaowen:Guten Nacht! Herr Dodson.
  Peter:Night. Thanks John.