Data Mining Methods and Models

DATA MINING METHODS AND MODELS......Page 3
 CONTENTS......Page 9
 PREFACE......Page 13
  Need for Dimension Reduction in Data Mining......Page 19
  Principal Components Analysis......Page 20
   Applying Principal Components Analysis to the Houses Data Set......Page 23
   How Many Components Should We Extract?......Page 27
   Profiling the Principal Components......Page 31
   Communalities......Page 33
   Validation of the Principal Components......Page 35
   Applying Factor Analysis to the Adult Data Set......Page 36
   Factor Rotation......Page 38
  User-Defined Composites......Page 41
   Example of a User-Defined Composite......Page 42
  Summary......Page 43
  Exercises......Page 46
 2 REGRESSION MODELING......Page 51
  Example of Simple Linear Regression......Page 52
  Least-Squares Estimates......Page 54
  Coefficient of Determination......Page 57
  Standard Error of the Estimate......Page 61
  Correlation Coefficient......Page 63
  ANOVA Table......Page 64
  Outliers, High Leverage Points, and Influential Observations......Page 66
  Regression Model......Page 73
  Inference in Regression......Page 75
   t-Test for the Relationship Between x and y......Page 76
   Confidence Interval for the Mean Value of y Given x......Page 78
   Prediction Interval for a Randomly Chosen Value of y Given x......Page 79
  Verifying the Regression Assumptions......Page 81
  Example: Baseball Data Set......Page 86
  Example: California Data Set......Page 92
  Transformations to Achieve Linearity......Page 97
   Box–Cox Transformations......Page 101
  Summary......Page 102
  Exercises......Page 104
  Example of Multiple Regression......Page 111
  Multiple Regression Model......Page 117
  Inference in Multiple Regression......Page 118
   t-Test for the Relationship Between y and x(i)......Page 119
   F-Test for the Significance of the Overall Regression Model......Page 120
   Confidence Interval for a Particular Coefficient......Page 122
  Regression with Categorical Predictors......Page 123
   Adjusting R(2): Penalizing Models for Including Predictors That Are Not Useful......Page 131
   Sequential Sums of Squares......Page 133
  Multicollinearity......Page 134
   Partial F-Test......Page 141
   Backward Elimination Procedure......Page 143
   All-Possible-Subsets Procedure......Page 144
   Forward Selection Procedure Applied to the Cereals Data Set......Page 145
   Backward Elimination Procedure Applied to the Cereals Data Set......Page 147
  Mallows’ C(p) Statistic......Page 149
  Variable Selection Criteria......Page 153
  Using the Principal Components as Predictors......Page 160
  Summary......Page 165
  Exercises......Page 167
 4 LOGISTIC REGRESSION......Page 173
  Simple Example of Logistic Regression......Page 174
  Maximum Likelihood Estimation......Page 176
  Interpreting Logistic Regression Output......Page 177
  Inference: Are the Predictors Significant?......Page 178
  Interpreting a Logistic Regression Model......Page 180
   Interpreting a Model for a Dichotomous Predictor......Page 181
   Interpreting a Model for a Polychotomous Predictor......Page 184
   Interpreting a Model for a Continuous Predictor......Page 188
  Assumption of Linearity......Page 192
  Zero-Cell Problem......Page 195
  Multiple Logistic Regression......Page 197
  Introducing Higher-Order Terms to Handle Nonlinearity......Page 201
  Validating the Logistic Regression Model......Page 207
  WEKA: Hands-on Analysis Using Logistic Regression......Page 212
  Summary......Page 215
  Exercises......Page 217
  Bayesian Approach......Page 222
  Maximum a Posteriori Classification......Page 224
   Posterior Odds Ratio......Page 228
   Balancing the Data......Page 230
  Naive Bayes Classification......Page 233
   Numeric Predictors......Page 237
  WEKA: Hands-on Analysis Using Naive Bayes......Page 241
   Clothing Purchase Example......Page 245
   Using the Bayesian Network to Find Probabilities......Page 247
  WEKA: Hands-On Analysis Using the Bayes Net Classifier......Page 250
  Summary......Page 252
  References......Page 254
  Exercises......Page 255
  Introduction to Genetic Algorithms......Page 258
  Basic Framework of a Genetic Algorithm......Page 259
  Simple Example of a Genetic Algorithm at Work......Page 261
  Modifications and Enhancements: Selection......Page 263
   Uniform Crossover......Page 265
   Simple Arithmetic Crossover......Page 266
  Using Genetic Algorithms to Train a Neural Network......Page 267
  WEKA: Hands-on Analysis Using Genetic Algorithms......Page 270
  Summary......Page 279
  References......Page 280
  Exercises......Page 281
  Cross-Industry Standard Process for Data Mining......Page 283
   Building the Cost/Benefit Table......Page 285
   Clothing Store Data Set......Page 288
   Transformations to Achieve Normality or Symmetry......Page 290
   Standardization and Flag Variables......Page 294
   Deriving New Variables......Page 295
   Exploring the Relationships Between the Predictors and the Response......Page 296
   Investigating the Correlation Structure Among the Predictors......Page 304
  Modeling and Evaluation Phases......Page 307
   Principal Components Analysis......Page 310
   Cluster Analysis: BIRCH Clustering Algorithm......Page 312
   Balancing the Training Data Set......Page 316
   Establishing the Baseline Model Performance......Page 317
   Model Collection A: Using the Principal Components......Page 318
   Overbalancing as a Surrogate for Misclassification Costs......Page 320
   Combining Models: Voting......Page 322
   Model Collection B: Non-PCA Models......Page 324
   Combining Models Using the Mean Response Probabilities......Page 326
  Summary......Page 330
  References......Page 334
 INDEX......Page 335