Homework Set 12, Due Wednesday October 26, 2016.
------------------------------------------------------

Assigned 10/17/2016, due 10/26


SIMULATION AND PARAMETER ESTIMATION FROM A MISSPECIFIED MODEL

Load the 199 scalar data values from the dataset "HW12dat.txt"
into your R workspace.

Problem: find the lognormal Maximum-Likelihood parameter estimates. 

Give their variances according to MLE theory, and compare the 
estimators of variance obtained by using the "sandwich" 
misspecified-model variance estimator.

NOTES: saying that the data Y_1,...,Y_n are LOGNORMAL is the same 
as saying that the transformed values  log(Y_i) are normally 
distributed with some unknown mean mu and variance sig^2 that 
must be estimated. You can obtain the closed form of the MLEs, 
but you are asked also to compute the gradients of the log-density 
terms in order to obtain the "sandwich" variance estimator.