Nonparametric Statistical Methods Using R  book cover
SAVE
$18.59
1st Edition

Nonparametric Statistical Methods Using R




ISBN 9781439873434
Published October 9, 2014 by Chapman and Hall/CRC
287 Pages - 58 B/W Illustrations

 
SAVE ~ $18.59
was $92.95
USD $74.36

Prices & shipping based on shipping country


Preview

Book Description

A Practical Guide to Implementing Nonparametric and Rank-Based Procedures

Nonparametric Statistical Methods Using R covers traditional nonparametric methods and rank-based analyses, including estimation and inference for models ranging from simple location models to general linear and nonlinear models for uncorrelated and correlated responses. The authors emphasize applications and statistical computation. They illustrate the methods with many real and simulated data examples using R, including the packages Rfit and npsm.

The book first gives an overview of the R language and basic statistical concepts before discussing nonparametrics. It presents rank-based methods for one- and two-sample problems, procedures for regression models, computation for general fixed-effects ANOVA and ANCOVA models, and time-to-event analyses. The last two chapters cover more advanced material, including high breakdown fits for general regression models and rank-based inference for cluster correlated data.

The book can be used as a primary text or supplement in a course on applied nonparametric or robust procedures and as a reference for researchers who need to implement nonparametric and rank-based methods in practice. Through numerous examples, it shows readers how to apply these methods using R.

Table of Contents

Getting Started with R
R Basics
Reading External Data
Generating Random Data
Graphics
Repeating Tasks
User-Defined Functions
Monte Carlo Simulation
R Packages

Basic Statistics
Sign Test
Signed-Rank Wilcoxon
Bootstrap
Robustness
One- and Two-Sample Proportion Problems
χ2 Tests

Two-Sample Problems
Introductory Example
Rank-Based Analyses
Scale Problem
Placement Test for the Behrens–Fisher Problem
Efficiency and Optimal Scores
Adaptive Rank Scores Tests

Regression I
Simple Linear Regression
Multiple Linear Regression
Linear Models
Aligned Rank Tests
Bootstrap
Nonparametric Regression
Correlation

ANOVA and ANCOVA
One-Way ANOVA
Multi-Way Crossed Factorial Design
ANCOVA
Methodology for Type III Hypotheses Testing
Ordered Alternatives
Multi-Sample Scale Problem

Time-to-Event Analysis
Kaplan–Meier and Log Rank Test
Cox Proportional Hazards Models
Accelerated Failure Time Models

Regression II
Robust Diagnostics
Weighted Regression
Linear Models with Skew Normal Errors
A Hogg-Type Adaptive Procedure
Nonlinear
Time Series

Cluster Correlated Data
Friedman’s Test
Joint Rankings Estimator
Robust Variance Component Estimators
Multiple Rankings Estimator
GEE-Type Estimator

Bibliography

Index

Exercises appear at the end of each chapter.

...
View More

Author(s)

Biography

John Kloke is a biostatistician and assistant scientist at the University of Wisconsin–Madison. He has held faculty positions at the University of Pittsburgh, Bucknell University, and Pomona College. An R user for more than 15 years, he is an author and maintainer of numerous R packages, including Rfit and npsm. He has published papers on nonparametric rank-based estimation, including analysis of cluster correlated data.

Joseph W. McKean is a professor of statistics at Western Michigan University. He has published many papers on nonparametric and robust statistical procedures and has co-authored several books, including Robust Nonparametric Statistical Methods and Introduction to Mathematical Statistics. He is an associate editor of several statistics journals and a fellow of the American Statistical Association.

Reviews

"In general, this textbook is a good addition to the sparse offerings in entry-level nonparametrics. This book would be especially good for the shelf of anyone who already knows nonparametrics, but wants a reference for how to apply those techniques in R. As R becomes more ubiquitous and data science grows into its own, I think this approach will become more common and this book will be shown to be ahead of its time." (The American Statistician)