Yes!

Yes.

The r function nlm minimizes arbitrary functions written in r. So to maximize the likelihood, we hand nlm the negative of the log likelihood.

Yes, the nlm function is widely used throughout the statistical community for this purpose.

You can cut and paste the following simple example. #=comment line.

# 1 Maximum likelihood in R (nlm)

# Using functions &

# numerical optimisation & the finding hessian

# numerically etc

#

# Recall the hessian is the negative of the observed

# information

#

# we use the non-linear minimization function in R - nlm

# which lives in the library(stats)

#

# earlier we did library (stats) # load library

?nlm # look at help for nlm

#

# Example 1: random sample of 100 Xs ~ N(100, 9)

#

myx

Yes, i've used. But especial interest must be taken with over/underflow.

Also, try the maxLik function (from the maxLik package). This is specifically designed for likelihood maximization.

please help me regarding this. what should I do it nlm replaces the values with maximum positive?

Hi

See the code in the previous answer I sent.

Copy the code into R and run it.

In the function:

loglikenorm2