An Analogy Permitting Maximum Likelihood Estimation by a Simple Modification of General Least Squares Algorithms
NAVAL MEDICAL RESEARCH INST BETHESDA MD
Pagination or Media Count:
The basis for a method of modifying general nonlinear least squares algorithms for maximum likelihood estimation is developed. The resultant algorithm provides a direct general approach to maximum likelihood estimation. Only the likelihood function for a single observation need be specified. Neither second partial derivatives nor weights are required. First partial derivatives may be obtained numerically and are used to compute the covariance matrix of the estimated parameters.
- Statistics and Probability