| |
| |
| Introduction | |
| |
| |
| What's in the Book | |
| |
| |
| Why Use Excel? | |
| |
| |
| |
| Components of Decision Analytics | |
| |
| |
| Classifying According to Existing Categories | |
| |
| |
| Using a Two-Step Approach | |
| |
| |
| Multiple Regression and Decision Analytics | |
| |
| |
| Access to a Reference Sample | |
| |
| |
| Multivariate Analysis of Variance | |
| |
| |
| Discriminant Function Analysis | |
| |
| |
| Logistic Regression | |
| |
| |
| Classifying According to Naturally Occurring Clusters | |
| |
| |
| Principal Components Analysis | |
| |
| |
| Cluster Analysis | |
| |
| |
| Some Terminology Problems | |
| |
| |
| The Design Sets the Terms | |
| |
| |
| Causation Versus Prediction | |
| |
| |
| Why the Terms Matter | |
| |
| |
| |
| Logistic Regression | |
| |
| |
| The Rationale for Logistic Regression | |
| |
| |
| The Scaling Problem | |
| |
| |
| About Underlying Assumptions | |
| |
| |
| Equal Spread | |
| |
| |
| Equal Variances with Dichotomies | |
| |
| |
| Equal Spread and the Range | |
| |
| |
| The Distribution of the Residuals | |
| |
| |
| Calculating the Residuals | |
| |
| |
| The Residuals of a Dichotomy | |
| |
| |
| Using Logistic Regression | |
| |
| |
| Using Odds Rather Than Probabilities | |
| |
| |
| Using Log Odds | |
| |
| |
| Using Maximum Likelihood Instead of Least Squares | |
| |
| |
| Maximizing the Log Likelihood | |
| |
| |
| Setting Up the Data | |
| |
| |
| Setting Up the Logistic Regression Equation | |
| |
| |
| Getting the Odds | |
| |
| |
| Getting the Probabilities | |
| |
| |
| Calculating the Log Likelihood | |
| |
| |
| Finding and Installing Solver | |
| |
| |
| Running Solver | |
| |
| |
| The Rationale for Log Likelihood | |
| |
| |
| The Probability of a Correct Classification | |
| |
| |
| Using the Log Likelihood | |
| |
| |
| The Statistical Significance of the Log Likelihood | |
| |
| |
| Setting Up the Reduced Model | |
| |
| |
| Setting Up the Full Model | |
| |
| |
| |
| Univariate Analysis of Variance (ANOVA) | |
| |
| |
| The Logic of ANOVA | |
| |
| |
| Using Variance | |
| |
| |
| Partitioning Variance | |
| |
| |
| Expected Values of Variances (Within Groups) | |
| |
| |
| Expected Values of Variances (Between Groups) | |
| |
| |
| The F-Ratio | |
| |
| |
| The Noncentral F Distribution | |
| |
| |
| Single Factor ANOVA | |
| |
| |
| Adopting an Error Rate | |
| |
| |
| Computing the Statistics | |
| |
| |
| Deriving the Standard Error of the Mean | |
| |
| |
| Using the Data Analysis Add-In | |
| |
| |
| Installing the Data Analysis Add-In | |
| |
| |
| Using the ANOVA: Single Factor Tool | |
| |
| |
| Understanding the ANOVA Output | |
| |
| |
| Using the Descriptive Statistics | |
| |
| |
| Using the Inferential Statistics | |
| |
| |
| The Regression Approach | |
| |
| |
| Using Effect Coding | |
| |
| |
| The LINEST() Formula | |
| |
| |
| The LINEST() Results | |
| |
| |
| LINEST() Inferential Statistics | |
| |
| |
| |
| Multivariate Analysis of Variance (MANOVA) | |
| |
| |
| The Rationale for MANOVA | |
| |
| |
| Correlated Variables | |
| |
| |
| Correlated Variables in ANOVA | |
| |
| |
| Visualizing Multivariate ANOVA | |
| |
| |
| Univariate ANOVA Results | |
| |
| |
| Multivariate ANOVA Results | |
| |
| |
| Means and Centroids | |
| |
| |
| From ANOVA to MANOVA | |
| |
| |
| Using SSCP Instead of SS | |
| |
| |
| Getting the Among and the Within SSCP Matrices | |
| |
| |
| Sums of Squares and SSCP Matrices | |
| |
| |
| Getting to a Multivariate F-Ratio | |
| |
| |
| Wilks' Lambda and the F-Ratio | |
| |
| |
| Converting Wilks' Lambda to an F Value | |
| |
| |
| Running a MANOVA in Excel | |
| |
| |
| Laying Out the Data | |
| |
| |
| Running the MANOVA Code | |
| |
| |
| Descriptive Statistics | |
| |
| |
| Equality of the Dispersion Matrices | |
| |
| |
| The Univariate and Multivariate F-Tests | |
| |
| |
| After the Multivariate Test | |
| |
| |
| |
| Discriminant Function Analysis: The Basics | |
| |
| |
| Treating a Category as a Number | |
| |
| |
| The Rationale for Discriminant Analysis | |
| |
| |
| Multiple Regression and Discriminant Analysis | |
| |
| |
| Adjusting Your Viewpoint | |
| |
| |
| Discriminant Analysis and Multiple Regression | |
| |
| |
| Regression, Discriminant Analysis, and Canonical Correlation | |
| |
| |
| Coding and Multiple Regression | |
| |
| |
| The Discriminant Function and the Regression Equation | |
| |
| |
| From Discriminant Weights to Regression Coefficients | |
| |
| |
| Eigenstructures from Regression and Discriminant Analysis | |
| |
| |
| Structure Coefficients Can Mislead | |
| |
| |
| Wrapping It Up | |
| |
| |
| |
| Discriminant Function Analysis: Further Issues | |
| |
| |
| Using the Discriminant Workbook | |
| |
| |
| Opening the Discriminant Workbook | |
| |
| |
| Using the Discriminant Dialog Box | |
| |
| |
| Why Run a Discriminant Analysis on Irises? | |
| |
| |
| Evaluating the Original Measures | |
| |
| |
| Discriminant Analysis and Investment | |
| |
| |
| Benchmarking with R | |
| |
| |
| Downloading R | |
| |
| |
| Arranging the Data File | |
| |
| |
| Running the Analysis | |
| |
| |
| The Results of the Discrim Add-In | |
| |
| |
| The Discriminant Results | |
| |
| |
| Interpreting the Structure Coefficients | |
| |
| |
| Eigenstructures and Coefficients | |
| |
| |
| Other Uses for the Coefficients | |
| |
| |
| Classifying the Cases | |
| |
| |
| Distance from the Centroids | |
| |
| |
| Correcting for the Means | |
| |
| |
| Adjusting for the Variance-Covariance Matrix | |
| |
| |
| Assigning a Classification | |
| |
| |
| Creating the Classification Table | |
| |
| |
| Training Samples: The Classification Is Known Beforehand | |
| |
| |
| |
| Principal Components Analysis | |
| |
| |
| Establishing a Conceptual Framework for Principal Components Analysis | |
| |
| |
| Principal Components and Tests | |
| |
| |
| PCA's Ground Rules | |
| |
| |
| Correlation and Oblique Factor Rotation | |
| |
| |
| Using the Principal Components Add-In | |
| |
| |
| The Correlation Matrix | |
| |
| |
| The Inverse of the R Matrix | |
| |
| |
| The Sphericity Test | |
| |
| |
| Counting Eigenvalues, Calculating Coefficients and Understanding Communalities | |
| |
| |
| How Many Components? | |
| |
| |
| Factor Score Coefficients | |
| |
| |
| Communalities | |
| |
| |
| Relationships Between the Individual Results | |
| |
| |
| Using the Eigenvalues and Eigenvectors | |
| |
| |
| Eigenvalues, Eigenvectors, and Loadings | |
| |
| |
| Eigenvalues, Eigenvectors, and Factor Coefficients | |
| |
| |
| Getting the Eigenvalues Directly from the Factor Scores | |
| |
| |
| Getting the Eigenvalues and Eigenvectors | |
| |
| |
| Iteration and Exhaustion | |
| |
| |
| Rotating Factors to a Meaningful Solution | |
| |
| |
| Identifying the Factors | |
| |
| |
| The Varimax Rotation | |
| |
| |
| Classification Examples | |
| |
| |
| State Crime Rates | |
| |
| |
| Physical Measurements of Aphids | |
| |
| |
| |
| Cluster Analysis: The Basics | |
| |
| |
| Cluster Analysis, Discriminant Analysis, and Logistic Regression | |
| |
| |
| Euclidean Distance | |
| |
| |
| Mahalanobis' D<sup>2</sup> and Cluster Analysis | |
| |
| |
| Finding Clusters: The Single Linkage Method | |
| |
| |
| The Self-Selecting Nature of Cluster Analysis | |
| |
| |
| Finding Clusters: The Complete Linkage Method | |
| |
| |
| Complete Linkage: An Example | |
| |
| |
| Other Linkage Methods | |
| |
| |
| Finding Clusters: The K-means Method | |
| |
| |
| Characteristics of K-means Analysis | |
| |
| |
| A K-means Example | |
| |
| |
| Benchmarking K-means with R | |
| |
| |
| |
| Cluster Analysis: Further Issues | |
| |
| |
| Using the K-means Workbook | |
| |
| |
| Deciding on the Number of Clusters | |
| |
| |
| The Cluster Members Worksheet | |
| |
| |
| The Cluster Centroids Worksheet | |
| |
| |
| The Cluster Variances Worksheet | |
| |
| |
| The F-Ratios Worksheet | |
| |
| |
| Reporting Process Statistics | |
| |
| |
| Cluster Analysis Using Principal Components | |
| |
| |
| Principal Components Revisited | |
| |
| |
| Clustering Wines | |
| |
| |
| Cross-Validating the Results | |
| |
| |
| Index | |