Preface | p. xiii |
Introduction | p. 1 |
Regression and Model Building | p. 1 |
Data Collection | p. 7 |
Uses of Regression | p. 11 |
Role of the Computer | p. 12 |
Simple Linear Regression | p. 13 |
Simple Linear Regression Model | p. 13 |
Least-Squares Estimation of the Parameters | p. 14 |
Estimation of [beta subscript 0] and [beta subscript 1] | p. 14 |
Properties of the Least-Squares Estimators and the Fitted Regression Model | p. 20 |
Estimation of [sigma superscript 2] | p. 22 |
An Alternate Form of the Model | p. 24 |
Hypothesis Testing on the Slope and Intercept | p. 24 |
Use of t-Tests | p. 25 |
Testing Significance of Regression | p. 26 |
The Analysis of Variance | p. 28 |
Interval Estimation in Simple Linear Regression | p. 32 |
Confidence Intervals on [beta subscript 0], [beta subscript 1], and [sigma superscript 2] | p. 32 |
Interval Estimation of the Mean Response | p. 34 |
Prediction of New Observations | p. 37 |
Coefficient of Determination | p. 39 |
Some Considerations in the Use of Regression | p. 41 |
Regression Through the Origin | p. 44 |
Estimation by Maximum Likelihood | p. 50 |
Case Where the Regressor x is Random | p. 52 |
x and y Jointly Distributed | p. 52 |
x and y Jointly Normally Distributed: The Correlation Model | p. 53 |
Problems | p. 58 |
Multiple Linear Regression | p. 67 |
Multiple Regression Models | p. 67 |
Estimation of the Model Parameters | p. 71 |
Least-Squares Estimation of the Regression Coefficients | p. 71 |
A Geometrical Interpretation of Least Squares | p. 81 |
Properties of the Least-Squares Estimators | p. 82 |
Estimation of [sigma superscript 2] | p. 82 |
Inadequacy of Scatter Diagrams in Multiple Regression | p. 84 |
Maximum-Likelihood Estimation | p. 85 |
Hypothesis Testing in Multiple Linear Regression | p. 87 |
Test for Significance of Regression | p. 87 |
Tests on Individual Regression Coefficients | p. 91 |
Special Case of Orthogonal Columns in X | p. 96 |
Testing the General Linear Hypothesis | p. 98 |
Confidence Intervals in Multiple Regression | p. 101 |
Confidence Intervals on the Regression Coefficients | p. 102 |
Confidence Interval Estimation of the Mean Response | p. 103 |
Simultaneous Confidence Intervals on Regression Coefficients | p. 104 |
Prediction of New Observations | p. 108 |
Hidden Extrapolation in Multiple Regression | p. 109 |
Standardized Regression Coefficients | p. 112 |
Multicollinearity | p. 117 |
Why Do Regression Coefficients have the Wrong Sign? | p. 120 |
Problems | p. 122 |
Model Adequacy Checking | p. 131 |
Introduction | p. 131 |
Residual Analysis | p. 132 |
Definition of Residuals | p. 132 |
Methods for Scaling Residuals | p. 132 |
Residual Plots | p. 138 |
Partial Regression and Partial Residual Plots | p. 146 |
Other Residual Plotting and Analysis Methods | p. 150 |
The PRESS Statistic | p. 152 |
Detection and Treatment of Outliers | p. 154 |
Lack of Fit of the Regression Model | p. 158 |
A Formal Test for Lack of Fit | p. 158 |
Estimation of Pure Error from Near-Neighbors | p. 162 |
Problems | p. 166 |
Transformations and Weighting to Correct Model Inadequacies | p. 173 |
Introduction | p. 173 |
Variance-Stabilizing Transformations | p. 174 |
Transformations to Linearize the Model | p. 178 |
Analytical Methods for Selecting a Transformation | p. 186 |
Transformations on y: The Box-Cox Method | p. 186 |
Transformations on the Regressor Variables | p. 189 |
Generalized and Weighted Least Squares | p. 193 |
Generalized Least Squares | p. 193 |
Weighted Least Squares | p. 195 |
Some Practical Issues | p. 196 |
Problems | p. 200 |
Diagnostics for Leverage and Influence | p. 207 |
Importance of Detecting Influential Observations | p. 207 |
Leverage | p. 209 |
Measures of Influence: Cook's D | p. 210 |
Measures of Influence: DFFITS and DFBETAS | p. 213 |
A Measure of Model Performance | p. 216 |
Detecting Groups of Influential Observations | p. 217 |
Treatment of Influential Observations | p. 218 |
Problems | p. 219 |
Polynomial Regression Models | p. 221 |
Introduction | p. 221 |
Polynomial Models in One Variable | p. 221 |
Basic Principles | p. 221 |
Piecewise Polynomial Fitting (Splines) | p. 228 |
Polynomial and Trigonometric Terms | p. 236 |
Nonparametric Regression | p. 237 |
Kernel Regression | p. 238 |
Locally Weighted Regression (Loess) | p. 239 |
Final Cautions | p. 243 |
Polynomial Models in Two or More Variables | p. 244 |
Orthogonal Polynomials | p. 253 |
Problems | p. 258 |
Indicator Variables | p. 265 |
The General Concept of Indicator Variables | p. 265 |
Comments on the Use of Indicator Variables | p. 279 |
Indicator Variables versus Regression on Allocated Codes | p. 279 |
Indicator Variables as a Substitute for a Quantitative Regressor | p. 280 |
Regression Approach to Analysis of Variance | p. 281 |
Problems | p. 287 |
Variable Selection and Model Building | p. 291 |
Introduction | p. 291 |
The Model-Building Problem | p. 291 |
Consequences of Model Misspecification | p. 292 |
Criteria for Evaluating Subset Regression Models | p. 296 |
Computational Techniques for Variable Selection | p. 302 |
All Possible Regressions | p. 302 |
Stepwise Regression Methods | p. 310 |
Some Final Recommendations for Practice | p. 317 |
Problems | p. 318 |
Multicollinearity | p. 325 |
Introduction | p. 325 |
Sources of Multicollinearity | p. 325 |
Effects of Multicollinearity | p. 328 |
Multicollinearity Diagnostics | p. 334 |
Examination of the Correlation Matrix | p. 334 |
Variance Inflation Factors | p. 337 |
Eigensystem Analysis of X'X | p. 339 |
Other Diagnostics | p. 343 |
Methods for Dealing with Multicollinearity | p. 345 |
Collecting Additional Data | p. 345 |
Model Respecification | p. 346 |
Ridge Regression | p. 348 |
Other Methods | p. 363 |
Comparison and Evaluation of Biased Estimators | p. 375 |
Problems | p. 378 |
Robust Regression | p. 382 |
The Need for Robust Regression | p. 382 |
M-Estimators | p. 386 |
Properties of Robust Estimators | p. 400 |
Breakdown Point | p. 400 |
Efficiency | p. 401 |
Survey of Other Robust Regression Estimators | p. 401 |
High-Breakdown-Point Estimators | p. 401 |
Bounded Influence Estimators | p. 406 |
Other Procedures | p. 407 |
Computing Robust Regression Estimators | p. 409 |
Problems | p. 410 |
Introduction to Nonlinear Regression | p. 414 |
Linear and Nonlinear Regression Models | p. 414 |
Linear Regression Models | p. 414 |
Nonlinear Regression Models | p. 415 |
Nonlinear Least Squares | p. 416 |
Transformation to a Linear Model | p. 420 |
Parameter Estimation in a Nonlinear System | p. 423 |
Linearization | p. 423 |
Other Parameter Estimation Methods | p. 431 |
Starting Values | p. 432 |
Computer Programs | p. 433 |
Statistical Inference in Nonlinear Regression | p. 434 |
Examples of Nonlinear Regression Models | p. 437 |
Problems | p. 438 |
Generalized Linear Models | p. 443 |
Introduction | p. 443 |
Logistic Regression Models | p. 444 |
Models with a Binary Response Variable | p. 444 |
Estimating the Parameters in a Logistic Regression Model | p. 447 |
Interpretation of the Parameters in a Logistic Regression Model | p. 450 |
Hypothesis Tests on Model Param6ters | p. 453 |
Poisson Regression | p. 459 |
The Generalized Linear Model | p. 466 |
Link Functions and Linear Predictors | p. 467 |
Parameter Estimation and Inference in the GLM | p. 468 |
Prediction and Estimation with the GLM | p. 472 |
Residual Analysis in the GLM | p. 474 |
Overdispersion | p. 475 |
Problems | p. 477 |
Other Topics in the Use of Regression Analysis | p. 488 |
Regression Models with Autocorrelation Errors | p. 488 |
Source and Effects of Autocorrelation | p. 488 |
Detecting the Presence of Autocorrelation | p. 489 |
Parameter Estimation Methods | p. 494 |
Effect of Measurement Errors in the Regressors | p. 500 |
Simple Linear Regression | p. 501 |
The Berkson Model | p. 502 |
Inverse Estimation--The Calibration Problem | p. 503 |
Bootstrapping in Regression | p. 508 |
Bootstrap Sampling in Regression | p. 509 |
Bootstrap Confidence Intervals | p. 510 |
Classification and Regression Trees (CART) | p. 516 |
Neural Networks | p. 518 |
Designed Experiments for Regression | p. 521 |
Problems | p. 524 |
Validation of Regression Models | p. 529 |
Introduction | p. 529 |
Validation Techniques | p. 530 |
Analysis of Model Coefficients and Predicted Values | p. 530 |
Collecting Fresh Data--Confirmation Runs | p. 532 |
Data Splitting | p. 534 |
Data from Planned Experiments | p. 545 |
Problems | p. 545 |
Statistical Tables | p. 549 |
Data Sets For Exercises | p. 567 |
Supplemental Technical Material | p. 582 |
Background on Basic Test Statistics | p. 582 |
Background from the Theory of Linear Models | p. 585 |
Important Results on SS[subscript R] and SS[subscript Res] | p. 588 |
The Gauss-Markov Theorem, Var([varepsilon]) = [sigma superscript 2]I | p. 594 |
Computational Aspects of Multiple Regression | p. 595 |
A Result on the Inverse of a Matrix | p. 597 |
Development of the PRESS Statistic | p. 598 |
Development of S[superscript 2 subscript (i)] | p. 600 |
An Outlier Test Based on R-Student | p. 601 |
The Gauss-Markov Theorem, Var([varepsilon]) = V | p. 604 |
The Bias in MS[subscript Res] When the Model is Underspecified | p. 606 |
Computation of Influence Diagnostics | p. 608 |
Generalized Linear Models | p. 610 |
References | p. 621 |
Index | p. 637 |
Table of Contents provided by Syndetics. All Rights Reserved. |