Linear models with R /
"Like its widely praised, best-selling predecessor, this edition combines statistics and R to seamlessly give a coherent exposition of the practice of linear modeling. The text offers up-to-date insight on essential data analysis topics, from estimation, inference, and prediction to missing dat...
Main Author: | |
---|---|
Format: | Book |
Language: | English |
Published: |
Boca Raton :
CRC Press, Taylor & Francis Group,
[2015]
Boca Raton : CRC Press, [2015] Boca Raton, FL : CRC Press/Taylor & Francis Group, [2015] Boca Raton, FL : [2015] |
Edition: | Second edition |
Series: | Texts in statistical science
|
Subjects: |
Table of Contents:
- 1. Introduction
- 2. Estimation
- 3. Inference
- 4. Prediction
- 5. Explanations
- 6. Diagnostics
- 7. Problems with the predictors
- 8. Problems with the errors
- 9. Transformation
- 10. Model selection
- 11. Shrinkage methods
- 12. Insurance redlining
- a complete example
- 13. Missing data
- 14. Categorical predictors
- 15. One factor models
- 16. Models with several factors
- 17. Experiments with blocks
- Appendix A. About R
- note: 1 Introduction
- 1.1. Before You Start
- 1.2. Initial Data Analysis
- 1.3. When to Use Linear Modeling
- 1.4. History
- 2. Estimation
- 2.1. Linear Model
- 2.2. Matrix Representation
- 2.3. Estimating β
- 2.4. Least Squares Estimation
- 2.5. Examples of Calculating β
- 2.6. Example
- 2.7. QR Decomposition
- 2.8. Gauss--Markov Theorem
- 2.9. Goodness of Fit
- 2.10. Identifiability
- 2.11. Orthogonality
- 3. Inference
- 3.1. Hypothesis Tests to Compare Models
- 3.2. Testing Examples
- 3.3. Permutation Tests
- 3.4. Sampling
- 3.5. Confidence Intervals for β
- 3.6. Bootstrap Confidence Intervals
- 4. Prediction
- 4.1. Confidence Intervals for Predictions
- 4.2. Predicting Body Fat
- 4.3. Autoregression
- 4.4. What Can Go Wrong with Predictions?
- 5. Explanation
- 5.1. Simple Meaning
- 5.2. Causality
- 5.3. Designed Experiments
- 5.4. Observational Data
- 5.5. Matching
- 5.6. Covariate Adjustment
- 5.7. Qualitative Support for Causation
- 6. Diagnostics
- 6.1. Checking Error Assumptions
- 6.1.1. Constant Variance
- 6.1.2. Normality
- 6.1.3. Correlated Errors
- 6.2. Finding Unusual Observations
- 6.2.1. Leverage
- 6.2.2. Outliers
- 6.2.3. Influential Observations
- 6.3. Checking the Structure of the Model
- 6.4. Discussion
- 7. Problems with the Predictors
- 7.1. Errors in the Predictors
- 7.2. Changes of Scale
- 7.3. Collinearity
- 8. Problems with the Error
- 8.1. Generalized Least Squares
- 8.2. Weighted Least Squares
- 8.3. Testing for Lack of Fit
- 8.4. Robust Regression
- 8.4.1. M-Estimation
- 8.4.2. Least Trimmed Squares
- 9. Transformation
- 9.1. Transforming the Response
- 9.2. Transforming the Predictors
- 9.3. Broken Stick Regression
- 9.4. Polynomials
- 9.5. Splines
- 9.6. Additive Models
- 9.7. More Complex Models
- 10. Model Selection
- 10.1. Hierarchical Models
- 10.2. Testing-Based Procedures
- 10.3. Criterion-Based Procedures
- 10.4. Summary
- 11. Shrinkage Methods
- 11.1. Principal Components
- 11.2. Partial Least Squares
- 11.3. Ridge Regression
- 11.4. Lasso
- 12. Insurance Redlining
- - A Complete Example
- 12.1. Ecological Correlation
- 12.2. Initial Data Analysis
- 12.3. Full Model and Diagnostics
- 12.4. Sensitivity Analysis
- 12.5. Discussion
- 13. Missing Data
- 13.1. Types of Missing Data
- 13.2. Deletion
- 13.3. Single Imputation
- 13.4. Multiple Imputation
- 14. Categorical Predictors
- 14.1. Two-Level Factor
- 14.2. Factors and Quantitative Predictors
- 14.3. Interpretation with Interaction Terms
- 14.4. Factors With More Than Two Levels
- 14.5. Alternative Codings of Qualitative Predictors
- 15. One Factor Models
- 15.1. Model
- 15.2. Example
- 15.3. Diagnostics
- 15.4. Pairwise Comparisons
- 15.5. False Discovery Rate
- 16. Models with Several Factors
- 16.1. Two Factors with No Replication
- 16.2. Two Factors with Replication
- 16.3. Two Factors with an Interaction
- 16.4. Larger Factorial Experiments
- 17. Experiments with Blocks
- 17.1. Randomized Block Design
- 17.2. Latin Squares
- 17.3. Balanced Incomplete Block Design.