The DMA Multiple Regression Equations transformation method shifts coordinate values between geodetic datums. It can be defined for local geodetic datums worldwide. The DMA Multiple Regression Equations method uses Doppler-derived parameters and provides a general solution with limited accuracy. It provides a transformation that is accurate to within 3-10 meters.
For a detailed discussion of the DMA Multiple Regression Equations algorithms and parameters for a variety of local geodetic datums, please refer to Defense Mapping Agency Technical Report TR 8350.2, 1991 "Department of Defense World Geodetic System 1984: Its Definition and Relationships with Local Geodetic Systems."
The main advantage of the DMA Multiple Regression Equations method lies in the modeling of distortions for datums that cover continental-sized land areas. This achieves a better fit in geodetic applications than the Molodensky method.
Note: The DMA Multiple Regression Equations method is an application of the theory of least squares. The coefficients for the mathematical regression equations are determined by fitting a polynomial to predicted shifts in a local area. If the DMA Multiple Regression Equations method is applied outside of the local area for which the coefficients of the equations are determined, the results may be unpredictable.