MLE of the (mu,var) parameter vector when we assume the sample is i.i.d. and drawn from the normal distribution.
Source:R/mle_normal.R
mle_normal.Rd
Of course, the draws are unlikely to be normal, but the normal distribution is often a good model. Hypothesis testing, such as relative likelihoods, can be used to assess the appropriateness of the normal model to the data.