#### SAS System for Regression,Third Edition

Author: **Rudolf Jakob Freund**

Learn to perform a wide variety of regression analyses using SAS software with this example-driven revised favorite from SAS Publishing. With this third edition you will learn the basics of performing regression analyses using a wide variety of models including nonlinear models. Other topics include performing linear regression analyses using PROVC REG and diagnosing and providing remedies for data problems, including outliers and multicollinearity. Examples feature numerous SAS procedures including REG, PLOT, GPLOT, NLIN, RSREG, AUTOREG, PRINCOMP, and others. A helpful discussion of theory is supplied where necessary. Some knowledge of both regression and the SAS System are assumed. The updated third edition includes revisions, updated material, and new material. You'll find information on using SAS/INSIGHT software, regression with a binary response with emphasis on PROC LOGISTIC, and nonparametric regression (smoothing) using moving averages and PROC LOESS. Additionally, updated material throughout the book includes high-resolution PROC REG graphics output, data sets by the OUTEST option described and illustrated, and using PROC SCORE to predict another data set. Supports releases 6.12 and higher of SAS software.

##### Table of Contents:

Acknowledgments | vii | |

Chapter 1 | Regression Concepts | 1 |

1.1 | Statistical Background | 1 |

1.2 | Performing a Regression with the IML Procedure | 9 |

1.3 | Regression with the SAS System | 12 |

Chapter 2 | Using the REG Procedure | 15 |

2.1 | Introduction | 15 |

2.2 | A Model with One Independent Variable | 17 |

2.3 | A Model with Several Independent Variables | 20 |

2.4 | Various MODEL Statement Options | 24 |

2.5 | Further Examination of Model Parameters | 36 |

2.6 | Plotting Results | 43 |

2.7 | Creating Data I: the OUTPUT Statement | 52 |

2.8 | Creating Data II: Other Data Sets | 55 |

2.9 | Creating Data III: ODS Output | 57 |

2.10 | Predicting to a Different Set of Data | 57 |

2.11 | Exact Collinearity: Linear Dependency | 60 |

2.12 | Summary | 62 |

Chapter 3 | Observations | 63 |

3.1 | Introduction | 63 |

3.2 | Outlier Detection | 64 |

3.3 | Specification Errors | 77 |

3.4 | Heterogeneous Variances | 81 |

3.5 | Correlated Errors | 86 |

3.6 | Summary | 94 |

Chapter 4 | Multicollinearity: Detection and Remedial Measures | 95 |

4.1 | Introduction | 95 |

4.2 | Detecting Multicollinearity | 97 |

4.3 | Model Restructuring | 101 |

4.4 | Variable Selection | 108 |

4.5 | Biased Estimation | 120 |

4.6 | Summary | 125 |

Chapter 5 | Curve Fitting | 127 |

5.1 | Introduction | 127 |

5.2 | Polynomial Models with One Independent Variable | 128 |

5.3 | Polynomial Plots | 132 |

5.4 | Polynomial Models with Several Variables | 134 |

5.5 | Response Surface Plots | 138 |

5.6 | A Three-Factor Response Surface Experiment | 142 |

5.7 | Curve Fitting without a Model | 148 |

5.8 | Summary | 155 |

Chapter 6 | Special Applications of Linear Models | 157 |

6.1 | Introduction | 157 |

6.2 | Multiplicative Models | 158 |

6.3 | Spline Models | 164 |

6.4 | Indicator Variables | 169 |

6.5 | Binary Response Variable: Logistic Regression | 173 |

6.6 | Summary | 184 |

Chapter 7 | Nonlinear Models | 185 |

7.1 | Introduction | 185 |

7.2 | Estimating the Exponential Decay Model | 186 |

7.3 | Fitting a Growth Curve with the NLIN Procedure | 192 |

7.4 | Fitting Splines with Unknown Knots | 198 |

7.5 | Additional Comments on the NLIN Procedure | 203 |

7.6 | Summary | 204 |

Chapter 8 | Using SAS/INSIGHT Software for Regression | 207 |

8.1 | Introduction | 207 |

8.2 | Multiple Linear Regression: the BOQ Data | 207 |

8.3 | A Polynomial Response Surface: the FISH Data | 215 |

8.4 | Logistic Regression: the DYSTRO Data | 218 |

8.5 | Nonparametric Smoothing: the Barbados Data | 221 |

8.6 | Summary | 225 |

References | 227 | |

Index | 229 |

New interesting textbook: **Indoor Grilling or White Wine For Dummies**

#### Optimal Control Systems

Author: **D Subbaram Naidu**

The theory of optimal control systems has grown and flourished since the 1960's. Many texts, written on varying levels of sophistication, have been published on the subject. Yet even those purportedly designed for beginners in the field are often riddled with complex theorems, and many treatments fail to include topics that are essential to a thorough grounding in the various aspects of and approaches to optimal control.

Optimal Control Systems provides a comprehensive but accessible treatment of the subject with just the right degree of mathematical rigor to be complete but practical. It provides a solid bridge between "traditional" optimization using the calculus of variations and what is called "modern" optimal control. It also treats both continuous-time and discrete-time optimal control systems, giving students a firm grasp on both methods. Among this book's most outstanding features is a summary table that accompanies each topic or problem and includes a statement of the problem with a step-by-step solution. Students will also gain valuable experience in using industry-standard MATLAB and SIMULINK software, including the Control System and Symbolic Math Toolboxes.

Diverse applications across fields from power engineering to medicine make a foundation in optimal control systems an essential part of an engineer's background. This clear, streamlined presentation is ideal for a graduate level course on control systems and as a quick reference for working engineers.

#### Booknews

Naidu (Idaho State University) introduces the theory and concepts useful for optimizing the control of physical systems which are dynamic and hence described by ordinary differential equations. Written for a second level graduate course, the textbook covers the calculus of variations, the linear quadratic regulator system, discrete-time systems, the Pontryagin minimum principle, and constraints on control and state variables. MATLAB and SIMULINK implementations are provided. Annotation c. Book News, Inc., Portland, OR

## No comments:

## Post a Comment