Maxlikmt provides methods for statistical inference of weighted or unweighted, bounded or unbounded maximum likelihood models.
- Inverted Hessian covariance matrix
- Heteroskedastic-consistent covariance matrix
- Wald confidence limits
- Likelihood ratio statistics
- Likelihood profile and Profile ‘t’ traces
- BFGS (Broyden, Fletcher, Goldfarb, and Powell)
- DFP (Davidon, Fletcher, and Powell)
- BHHH (Berndt, Hall, Hall, and Hausman)
Line search methods
- Brent’s method
- Strong Wolfe’s Conditions
- Bounded parameters.
- Specify fixed and free parameters.
- Dynamic algorithm switching.
- Compute all, a subset, or none of the derivatives numerically.
- Easily pass data other than the model parameters as extra input arguments. New!
- Methods to simply create matrices needed for log-likelihood computation from subsets of the parameters.
- Threaded and thread-safe
- Option to avoid computations that are the same for the log-likelihood function and derivatives.
- The tremendous speed of user-defined procedures in GAUSS speeds up your estimation.
For more than 30 years, leading researchers have trusted the efficient and numerically sound code in the GAUSS maximum likelihood estimation tools to keep them at the forefront of their fields.
Novice users will typically leave most of these options at the default values. However, they can be a great help when tackling more difficult problems.
|Simple parameter bounds of the type: lower_bd ≤ x_i ≤ upper_bd
|BFGS, DFP, Newton and BHHH.
|Specify descent algorithms to switch between based upon the number of elapsed iterations, a minimum change in the objective function or line search step size.
|Covariance matrix type
|Compute a ML covariance matrix, a QML covariance matrix, or none.
|Probability level for statistical tests.
|Line search method
|STEPBT (quadratic and cubic curve fit), Brent’s method, BHHHStep, half-step or Strong Wolfe’s Conditions.
|Control which parameters are active (to be estimated) and which should be fixed to their start value.
|Either compute an analytical gradient, or have Maxlikmt compute a numerical gradient using the forward, central or backwards difference method.
|Either compute an analytical Hessian, or have Maxlikmt compute a numerical Hessian using the forward, central or backwards difference method.
|Compares the analytical gradient computed by the user supplied function with the numerical gradient to check the analytical gradient for correctness.
|Starting seed value used by the random line search method to allow for repeatable code.
|Controls whether (or how often) iteration output is printed and whether a final report is printed.
|Advanced feature: Controls the increment size for computing the step size for numerical first and second derivatives.
|Random search radius
|The radius of the random search if attempted.
|Maximum iterations to converge.
|Maximum elapsed time
|Maximum number of minutes to converge.
|Maximum random search attempts
|Maximum allowed number of random line search attempts.
|Convergence is achieved when the direction vector changes less than this amount.