
Accession Number : ADA195603
Title : Complete Pivotal Gaussian Elimination Using the Veterans' Administration File Manager.
Descriptive Note : Final rept.,
Corporate Author : NAVAL HEALTH RESEARCH CENTER SAN DIEGO CA
Personal Author(s) : Hodgins, Dallas R
PDF Url : ADA195603
Report Date : 16 Mar 1988
Pagination or Media Count : 27
Abstract : The choice of a programming language in research and development is a serious commitment of time, money, and personnel. The principal question addressed is whether MUMPS is a viable language for developing statistical models that require numerical analysis. An algorithm to solve the linear equations of multiple regression is developed to explore and exploit the characteristics of the MUMPS language in handling numbers. Also in interest is the selection of a database management system and the conceptual schema of the data structure. Files structured relationally in the Veterans' Administration File Manager are manipulated algebraically. The development of concise, efficient language constructs lend support to using MUMPS is small sample analysis in statistical research. At one level, statistical analysis deals with discrete values in a twodimensional research. Assigning values to the points of this space is the equivalent of a function over the space. If one has a function (coupled pairs), the application of set theory is obvious. To most efficiently use set theory and the power of the notation of linear algebra, the data should be structured relationally. The MUMPS language, is conjunction with the Veteran's Administration File Manager used to create relational data structures, is a powerful, versatile media for numerical analysis. Using the notation of linear algebra and properly exploiting the inherent data storage virtues of MUMPS leads to efficient, fast code as is exemplified in the algorithm for complete pivotal forward Gaussian elimination presented in this paper. (sdw)
Descriptors : *DATA MANAGEMENT, *STATISTICAL ANALYSIS, *VETERANS(MILITARY PERSONNEL), *COMPUTER FILES, ALGORITHMS, COUPLING(INTERACTION), DATA BASES, DATA STORAGE SYSTEMS, EFFICIENCY, HANDLING, LANGUAGE, LINEAR ALGEBRA, LINEAR ALGEBRAIC EQUATIONS, MATHEMATICAL MODELS, NUMBERS, NUMERICAL ANALYSIS, PROGRAMMING LANGUAGES, REGRESSION ANALYSIS, SET THEORY, STATISTICS, TWO DIMENSIONAL, VIABILITY
Subject Categories : Statistics and Probability
Computer Programming and Software
Distribution Statement : APPROVED FOR PUBLIC RELEASE