Book Selection
- PDF / 89,731 Bytes
- 4 Pages / 595 x 794 pts Page_size
- 96 Downloads / 175 Views
r 2003 Operational Research Society Ltd. All rights reserved. 0160-5682/03 $25.00 www.palgrave-journals.com/jors
Book Selection Edited by J Crocker ZA Karian and EJ Dudewicz: Fitting Statistical Distributions: The Generalized Lambda Distribution and Generalized Bootstrap Methods I Nonako and D Teece: Managing Industrial Knowledge: Creation, Transfer and Utilization S Gass and C Harris: Encyclopedia of Operations Research and Management Science: Second ‘Centennial’ Edition D Sherwood: Smart Things to Know About Innovation and Creativity
Fitting Statistical Distributions: The Generalized Lambda Distribution and Generalized Bootstrap Methods ZA Karian and EJ Dudewicz Chapman & Hall/CRC Press, 2000. xvii þ 438pp. d53.99 ISBN: 1-58488-069-4 Much of the basic bivariate theory included here dates back to Plackett (1965) and Pearson (1913). In recent years, the authors have developed and extended the univariate, four parameter, generalised lambda distribution (GLD) approach due to Ramberg and Schmeiser. That approach dates from the 1970s. Fits, based on the tables provided, to a wide range of wellknown distributions and a small collection of data sets are included. The fits are generally very good. The univariate fitting family, the extended generalised lambda distribution (EGLD), presented in this book is very rich due to: (i) a recent (1999) percentile-based approach to fitting the GLD which complements the basic first-four-momentsmatching-based one; (ii) the use of a, scaled and shifted, generalised beta distribution (GBD) to complement the GLD family. In Chapter 1, the GLD family is defined through the inverse distribution function or percentile function. Thus, random variable realisations for a simulation project can easily be generated from any distribution that can be modelled by a GLD. The first parameter is a simple shift parameter. The second parameter is a scale parameter. Much of the analysis in this chapter is concerned with valid parameter values for the last two parameters and the associated shape and support implications. Chapter 2 shows how to estimate the (four) unknown parameters with the method of moments. The values of the last two parameters also determine whether the kth moments
671 672 672 673
of a GLD exist. When the first four moments of a GLD exist, a closed form for those moments is available which allows the method of moments to be applied. While Maple programs are provided, which can produce multiple ‘solutions’, many OR practitioners will be content to use the tables provided for the estimation of the four parameters. The EGLD is developed to cover all possible moments in Chapter 3. This solves limited coverage of moment–space problem found in some GLD applications. This is done by augmenting the GLD with a shifted and scaled, GBD that covers the skewness and kurtosis values not covered by the GLD. There is some important overlap between GBD and GLD, for example, for N(0, l). Both GBD and GLD are four-parameter families. The EGLD approach, due to the book’s authors, dates from 1996. Chapter 4 prov
Data Loading...
