I have a question related to: Why is a likelihood-ratio test distributed chi-squared?
On point 2 of @StasK's answer, he states:
The theorem assumes that all the relevant derivatives are non-zero. This can be challenged with some nonlinear problems and/or parameterizations, and/or situations when a parameter is not identified under the null.
- What derivatives are these?
- In the case of mixture distributions, when a parameter is not identified under the null, is it possible for the Information matrix to not be invertible? How?