maximum likelihood
(1.4 hours to learn)
Summary
Maximum likelihood is a general and powerful technique for learning statistical models, i.e. fitting the parameters to data. The maximum likelihood parameters are the ones under which the observed data has the highest probability. It is widely used in practice, and techniques such as Bayesian parameter estimation are closely related to maximum likelihood.
Context
This concept has the prerequisites:
- random variables
- independent random variables (The data are generally assumed to be independent draws from a distribution.)
- optimization problems (Maximum likelihood is formulated as an optimization problem.)
- Gaussian distribution (Fitting a Gaussian distribution is an instructive example of maximum likelihood estimation.)
See also
-No Additional Notes-