1st Edition

Statistical Simulation Power Method Polynomials and Other Transformations

By Todd C. Headrick Copyright 2010
    174 Pages 6 B/W Illustrations
    by Chapman & Hall

    174 Pages 6 B/W Illustrations
    by Chapman & Hall

    Although power method polynomials based on the standard normal distributions have been used in many different contexts for the past 30 years, it was not until recently that the probability density function (pdf) and cumulative distribution function (cdf) were derived and made available. Focusing on both univariate and multivariate nonnormal data generation, Statistical Simulation: Power Method Polynomials and Other Transformations presents techniques for conducting a Monte Carlo simulation study. It shows how to use power method polynomials for simulating univariate and multivariate nonnormal distributions with specified cumulants and correlation matrices.

    The book first explores the methodology underlying the power method, before demonstrating this method through examples of standard normal, logistic, and uniform power method pdfs. It also discusses methods for improving the performance of a simulation based on power method polynomials. The book then develops simulation procedures for systems of linear statistical models, intraclass correlation coefficients, and correlated continuous variates and ranks. Numerical examples and results from Monte Carlo simulations illustrate these procedures. The final chapter describes how the g-and-h and generalized lambda distribution (GLD) transformations are special applications of the more general multivariate nonnormal data generation approach. Throughout the text, the author employs Mathematica® in a range of procedures and offers the source code for download online.

    Written by a longtime researcher of the power method, this book explains how to simulate nonnormal distributions via easy-to-use power method polynomials. By using the methodology and techniques developed in the text, readers can evaluate different transformations in terms of comparing percentiles, measures of central tendency, goodness-of-fit tests, and more.

    Introduction

    The Power Method Transformation

    Univariate Theory

    Third-Order Systems

    Fifth-Order Systems

    Mathematica® Functions

    Limitations

    Multivariate Theory

    Using the Power Method Transformation

    Introduction

    Examples of Third- and Fifth-Order Polynomials

    Remediation Techniques

    Monte Carlo Simulation

    Some Further Considerations

    Simulating More Elaborate Correlation Structures

    Introduction

    Simulating Systems of Linear Statistical Models

    Methodology

    Numerical Example and Monte Carlo Simulation

    Some Additional Comments

    Simulating Intraclass Correlation Coefficients

    Methodology

    Numerical Example and Monte Carlo Simulation

    Simulating Correlated Continuous Variates and Ranks

    Methodology

    Numerical Example and Monte Carlo Simulation

    Some Additional Comments

    Other Transformations: The g-and-h and GLD Families of Distributions

    Introduction

    The g-and-h Family

    The Generalized Lambda Distributions (GLDs)

    Numerical Examples

    Multivariate Data Generation

    References

    Index

    Biography

    Todd C. Headrick is an associate professor and coordinator of the Educational Statistics & Measurement program at Southern Illinois University Carbondale.

    Headrick’s book is a valuable addition to the simulation world in the sense that it is the first systematic book on power method polynomials … this book has potential to be one of the key sources for researchers who are involved in random number generation and Monte Carlo studies. … it can serve and seems to be designed as a supplemental text in graduate level simulation- and computation-oriented courses. … Mathematica functions are provided throughout the book, which is a nice feature … Headrick’s book is a fairly major contribution to the literature … . I would highly recommend this book to anyone who deals with random number generation and more generally with Monte Carlo simulation and statistical computing.
    Journal of Statistical Software, Vol. 43, September 2011

    This interesting monograph concerns the use of power method polynomials in the context of simulating univariate and multivariate nonnormal distributions with given cumulant and correlation structure. Its strength is in the fact that the power method, g-and-h transformations and GLD transformations are capable of simulation multivariate nonnormal continuous distributions specified by its cumulants and correlation matrices. … —Zentralblatt MATH 1187