Basic Statistics Using SAS Enterprise Guide, a Primer by G. Der; B. S. Everitt Review by: Alex Karagrigoriou Journal of the Royal Statistical Society. Series A (Statistics in Society), Vol. 172, No. 2 (Apr., 2009), pp. 530-531 Published by: Wiley for the Royal Statistical Society Stable URL: http://www.jstor.org/stable/20622514 . Accessed: 28/06/2014 16:52 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact
[email protected]. . Wiley and Royal Statistical Society are collaborating with JSTOR to digitize, preserve and extend access to Journal of the Royal Statistical Society. Series A (Statistics in Society). http://www.jstor.org This content downloaded from 193.142.30.55 on Sat, 28 Jun 2014 16:52:44 PM All use subject to JSTOR Terms and Conditions http://www.jstor.org/action/showPublisher?publisherCode=black http://www.jstor.org/action/showPublisher?publisherCode=rss http://www.jstor.org/stable/20622514?origin=JSTOR-pdf http://www.jstor.org/page/info/about/policies/terms.jsp http://www.jstor.org/page/info/about/policies/terms.jsp 530 Book reviews Statistical and Probabilistic Methods in Actuarial Science P. J. Boland, 2007 Boca Raton, Chapman and Hall-CRC xvi + 352 pp., $79.95 ISBN 978-1-584-88695-2 Published under the 'Interdisciplinary statistics' series, the book under review has emerged from the lecture notes of the author. The book discusses a variety of practical issues in insurance, actuarial science and finance. Prerequisite for reading the book is a basic knowledge of probability and sta tistics. The book has an introduction followed by contents, eight chapters, a list of references, four appendices and an index. Every chapter ends with a large list of problems, some of which are theoreti cal and the rest are data analytic and application oriented. The appendices contain basics of proba bility and statistics, probability distributions, Bayes ian statistics and answers to selected problems. References, though not exhaustive, give a flavour of books for additional reading. The first chapter deals with the methodology of setting aside reserves to handle claims in insurance. After discussing the way that claims and reserves evolve, chain ladder meth ods, the average cost per claim method and loss ratio method for estimating reserves are discussed. Chap ter 2 introduces loss distributions such as the expo nential, Weibull, gamma, Pareto and log-normal. Chapter 3 is titled 'Risk theory' and discusses risk modelling by using compound distributions. The fourth chapter is on the probability of ruin in a sur plus process. Chapter 5 deals with credibility theory, which includes the Bayesian approach, and Chap ter 6 is on no-claim discounting in motor insurance. Chapter 7 is on generalized linear models and the last chapter is on decision and game theory, which is useful to insurance strategists. Examples given in all the chapters help to understand the concepts bet ter. There are very few typographical errors in the book. The book is a welcome addition to the vast lit erature on mathematical methods in insurance and readers with a little background in probability and statistics will benefit greatly from this book. Sreenivasan Ravi University of Mysore E-mail:
[email protected] Semi-supervised Learning O. Chappelle, B. Sch?lkopf and A. Zien (eds), 2006 Cambridge, MIT Press 508 pp., ?33.95 ISBN 978-0-262-03358-9 Machine learning distinguishes supervised tasks, that involve labelled or marked data (e.g. regres sion and classification), from unsupervised tasks, that involve unlabelled data (e.g. clustering, den sity estimation and principal component analysis). Thanks to advances in data collection and storage, an increasing number of application areas provide a mixture of labelled and unlabelled data, and the semisupervised problem involves using both types of data for a variety of inference tasks. This book provides a thorough overview of semisupervising learning, including taxonomy, al gorithms and applications. The brief introductory chapter introduces a pleasing analysis of when semi supervised learning is expected to work. The book is carefully structured, primarily according to the classes of algorithmic procedure that are available. These procedures correspond to very different types of representation, including probability densities and graphs. Ideas from Bayesian inference and sup port vector machines are frequently explored. Some practical and application-oriented chap ters are included. Of particular merit is an explor ation of semisupervised algorithms for large-scale applications, and a benchmarking study. The final part of the book is more reflective on future prospects and links with other parts of machine learning. Overall, this is a very well-structured and nicely prepared book, that provides good coverage of this emerging field. Niall Adams Imperial College London E-mail:
[email protected] Basic Statistics using SAS Enterprise Guide, a Primer G. Der and B. S. Everitt, 2007 Cary, SAS Institute Press 244 pp., $39.95 ISBN 978-1-599-94573-6 The SAS Enterprise Guide (EG) is a convenient point-and-click interface to the SAS system. It can be considered as the Excel of the SAS system since it allows you to access and manipulate data and at the same time to execute SAS programs and to produce statistical analysis results. It should be mentioned that the SAS programming language is not neces sary to be known thoroughly. Furthermore, the EG operations are performed quickly and the SAS pro gramming code is generated automatically for the user. This book describes the use of the SAS EG for basic statistical analysis. The first chapter is an introductory one which focuses on the creation and This content downloaded from 193.142.30.55 on Sat, 28 Jun 2014 16:52:44 PM All use subject to JSTOR Terms and Conditions http://www.jstor.org/page/info/about/policies/terms.jsp Book reviews 531 maintenance of 'projects' through which the EG organizes all aspects of statistical analysis (data, tasks and results). Although the EG interface pro vides four visible windows, statistical analyses (like those in this book) rely (aside from the menu bar) almost exclusively on one window only, namely the 'Project designer window'. The chapter deals also with ways to import, manipulate, organize and store data sets. The chapter ends with a brief discussion on the analysis tasks which can be found under the 'Analyze' menu, although the 'Describe' and the 'Data' menus are also frequently used for perform ing such tasks. Chapter 2 deals with numerous basic statistical topics and describes summary statistics (means, variances, etc.), basic graphs (histograms, boxplots, etc.) and the Student Mest along with the corres ponding non-parametric Wilcoxon test. The good ness-of-fit tests for categorical data are the main focus of Chapter 3, which covers the standard x2 test and its alternatives for low counts (Fisher's exact test) and for matched pairs data (McNemar's test). Chapter 4 deals with the analysis of bivariate data and covers scatter plots, correlation coefficients and simple linear regression analysis. Chapter 5 focuses on analysis of variance including factorial designs and Scheffe's method for multiple compar isons whereas Chapter 6 focuses on multiple linear regression including regression diagnostics. Logistic regression is the main theme of Chap ter 7 which includes a section on interpreting the estimated regression coefficients of the model and another on discussing the odds and posing (and answering) the question 'what are odds?'. Chapter 8 covers survival analysis problems. The Kaplan-Meier estimate of the survival function, the hazard function and the log-rank test for the equal ity of two survival functions are presented. The chapter ends with the Cox regression model. It is important to mention that the material of each chapter is presented with the help of a real example and each chapter ends with the analysis of at least one (usually two) more real example. One of the advantages of the book is the explana tions that are given occasionally for various delicate statistical terms (e.g. 'What is a /rvalue?' on page 41, 'What are odds?' on page 174 and 'How is the log odds value interpreted?' on page 179). Further more, the author provides short theoretical discus sions for various statistical concepts that are useful to the reader (e.g. type I versus type III sums of squares in Chapter 5). The use of a specific example to present the mate rial makes the reader comprehend much better the capabilities of the EG. In some cases, though, no reference is provided for the data sets that are used. Also at the end of each chapter more exercises or rather more data sets, preferably real, could have been provided. Extremely helpful the example programs and data sets of the book are accessible through the SAS Web site. In summary, the book is well written, easy to read and comprehend and serves nicely the purpose for which it was written, namely to introduce and at tract attention to SAS EG. The author has definitely succeeded by showing that SAS EG is easy to use and convenient and enables anyone to perform sta tistical analyses without having to write programs and code. The book is well suited for students, pro fessionals and researchers irrespective of their level of familiarity with SAS. Alex Karagrigoriou University of Cyprus Nicosia E-mail:
[email protected] Medical Biostatistics, 2nd edn A. Indrayan, 2008 Boca Raton, Chapman and Hall-CRC xlvi + 772 pp., $99.95 ISBN 978-1-584-88887-1 As outlined by the author in his objective, this book covers a very broad range of topics. The stated aim is to guide professionals such as epidemiologists, nurses, nutritionists, medical technicians, resear chers and health practitioners in the understand ing of medical biostatistics. The statistical concepts have been embedded in highly medically related examples and applications that have probably made the book ideal for the medically qualified reader, although for others this may be troublesome. Very basic aspects of statistics, e.g. Gaussian, Bernoulli and binomial distributions, as well as more advanced methods are presented. The latter include simple linear regression models and several variations of these, e.g. analysis-of-variance and multivariate analysis-of-variance methods, non parametric tests, logistic regression and survival analysis. There are introductory chapters on basic methodological topics such as study design, sam pling methods, clinical trials, probability laws, mean and variation measurements, confidence intervals, /?-values and Bayes's rule. Hence the book covers material that is usually within basic medical statis tics texts but expands also way beyond these. The great volume of applications, examples and data analyses that it includes has been successfully sum marized in a highly useful table at the start of the book, referring the reader to the appropriate section for different data analysis scenarios. This content downloaded from 193.142.30.55 on Sat, 28 Jun 2014 16:52:44 PM All use subject to JSTOR Terms and Conditions http://www.jstor.org/page/info/about/policies/terms.jsp Article Contents p. 530 p. 531 Issue Table of Contents Journal of the Royal Statistical Society. Series A (Statistics in Society), Vol. 172, No. 2 (Apr., 2009), pp. 287-534 Front Matter Modern Statistics: The Myth and the Magic [pp. 287-306] Overeducation and the Skills of UK Graduates [pp. 307-337] Surveying Migrant Households: A Comparison of Census-Based, Snowball and Intercept Point Surveys [pp. 339-360] Multilevel Modelling of Refusal and Non-Contact in Household Surveys: Evidence from Six UK Government Surveys [pp. 361-381] Accounting for Uncertainty in Health Economic Decision Models by Using Model Averaging [pp. 383-404] Modelling Population-Based Cancer Survival Trends by Using Join Point Models for Grouped Survival Data [pp. 405-425] Friendship Ties and Geographical Mobility: Evidence from Great Britain [pp. 427-442] Analysing Direct Effects in Randomized Trials with Secondary Interventions: An Application to Human Immunodeficiency Virus Prevention Trials [pp. 443-465] Sir Godfrey Thomson: A Statistical Pioneer [pp. 467-482] On Square Ordinal Contingency Tables: A Comparison of Social Class and Income Mobility for the Same Individuals [pp. 483-493] Cluster Designs to Assess the Prevalence of Acute Malnutrition by Lot Quality Assurance Sampling: A Validation Study by Computer Simulation [pp. 495-510] Multiple Imputation for Combining Confidential Data Owned by Two Agencies [pp. 511-528] Book Reviews Review: untitled [pp. 530-530] Review: untitled [pp. 530-530] Review: untitled [pp. 530-531] Review: untitled [pp. 531-532] Review: untitled [pp. 532-532] Review: untitled [pp. 532-533] Review: untitled [pp. 533-534] Back Matter