eBook

- Explores the well-known methods of OLS and maximum likelihood regression
- Develops specialized regression techniques, including nonparametric, logistic, Bayesian, robust, fuzzy, spatial, polynomial, and more
- Covers nonlinear and time series modeling
- Explains the output of SAS programs
- Offers SAS code for readers to immediately solve problems
- Confines proofs and derivations to the appendices, making the main presentation streamlined

**Regression Modeling: Methods, Theory, and Computation with SAS** provides an introduction to a diverse assortment of regression techniques using SAS to solve a wide variety of regression problems. The author fully documents the SAS programs and thoroughly explains the output produced by the programs.

The text presents the popular ordinary least squares (OLS) approach before introducing many alternative regression methods. It covers nonparametric regression, logistic regression (including Poisson regression), Bayesian regression, robust regression, fuzzy regression, random coefficients regression, *L*_{1} and *q*-quantile regression, regression in a spatial domain, ridge regression, semiparametric regression, nonlinear least squares, and time-series regression issues. For most of the regression methods, the author includes SAS procedure code, enabling readers to promptly perform their own regression runs.

*A Comprehensive, Accessible Source on Regression Methodology and Modeling*Requiring only basic knowledge of statistics and calculus, this book discusses how to use regression analysis for decision making and problem solving. It shows readers the power and diversity of regression techniques without overwhelming them with calculations.

Preface. Review of Fundamentals of Statistics. Bivariate Linear Regression and Correlation. Misspecified Disturbance Terms. Nonparametric Regression. Logistic Regression. Bayesian Regression. Robust Regression. Fuzzy Regression. Random Coefficients Regression.* L*_{1} and *q*-Quantile Regression. Regression in a Spatial Domain. Multiple Regression. Normal Correlation Models. Ridge Regression. Indicator Variables. Polynomial Model Estimation. Semiparametric Regression. Nonlinear Regression. Issues in Time Series Modeling and Estimation. Appendix. References. Index.

This highly readable book should be useful for students, lecturers, and practitioners alike as it covers most of the standard regression techniques and even some methods beyond.

—Karsten Webel, *Statistical Papers* (2012) 53

As an introductory text, it is mostly successful … . One of the great strengths of the text is that the examples tend to be linked into a structure … so that a student can more easily see how each procedure is connected to the concepts that preceded it. Another strength of the book is the detailed appendices at the end of each chapter. … A unique feature of this book is that it contains many chapters on facets of regression which are not covered in typical introductory texts … the book has great expository strength. It contains detailed verbal descriptions of the procedures used and the reasoning behind them, and these are always clear and linked to the previous descriptions. … the book serves as an excellent conceptual aid to a professor who would prefer to emphasize statistical reasoning to students, rather than to just rely upon the formulaic structure.

—*The American Statistician*, November 2010, Vol. 64, No. 4

In his book, Michael Panik takes up many aspects of modeling with a pedagogical approach, helping the reader to understand the process of the problem and proposed methods. The appendices enrich his process to [readers] who want to increase their knowledge. … this book is a very good tool for students and teachers in statistics, but also for researchers wishing to improve their knowledge in statistical modeling to apply it in their expertise domain.

—Christian Derquenne, *Journal of Statistical Software*, February 2010