Berndt–Hall–Hall–Hausman algorithm

The Berndt–Hall–Hall–Hausman (BHHH) algorithm is a numerical optimization algorithm similar to the Gauss–Newton algorithm. It is named after the four originators: Ernst R. Berndt, B. Hall, Robert Hall, and Jerry Hausman.

Usage

If a nonlinear model is fitted to the data one often needs to estimate coefficients through optimization. A number of optimisation algorithms have the following general structure. Suppose that the function to be optimized is Q(β). Then the algorithms are iterative, defining a sequence of approximations, β_k given by

$\beta_{k+1}=\beta_{k}-\lambda_{k}A_{k}\frac{\partial Q}{\partial \beta}(\beta_{k}),$,

where $\beta_{k}$ is the parameter estimate at step k, and $\lambda_{k}$ is a parameter (called step size) which partly determines the particular algorithm. For the BHHH algorithm λ_k is determined by calculations within a given iterative step, involving a line-search until a point β_k+1 is found satisfying certain criteria. In addition, for the BHHH algorithm, Q has the form

$Q = \sum_{i=1}^{N} Q_i$

and A is calculated using

$A_{k}=\left[\sum_{i=1}^{N}\frac{\partial \ln Q_i}{\partial \beta}(\beta_{k})\frac{\partial \ln Q_i}{\partial \beta}(\beta_{k})'\right]^{-1} .$

In other cases, e.g. Newton–Raphson, $A_{k}$ can have other forms. The BHHH algorithm has the advantage that, if certain conditions apply, convergence of the iterative procedure is guaranteed.

See also

• Davidon–Fletcher–Powell (DFP) algorithm
• Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm