Data Mining with SPSS Modeler

Transcription

1 Tilo Wendler Soren Grottrup Data Mining with SPSS Modeler Theory, Exercises and Solutions Springer

2 1 Introduction The Concept of the SPSS Modeler Structure and Features of This Book Prerequisites for Using This Book Structure of the Book and the Exercise/Solution Concept Using the Data and Streams Provided with the Book Datasets Provided with This Book Template Concept of This Book Introducing the Modeling Process Exercises Solutions IB Literature 22 2 Basic Functions of the SPSS Modeler Defining Streams and Scrolling Through a Dataset Switching Between Different Streams Defining or Modifying Value Labels Adding Comments to a Stream Exercises Solutions Data Handling and Sampling Methods Theory Calculations String Functions Extracting/Selecting Records Filtering Data Data Standardization: Z-Transformation Partitioning Datasets Sampling Methods Merge Datasets Ill vii

3 viii Contents Append Datasets Exercises Solutions 147 Literature Univariate Statistics Theory Discrete Versus Continuous Variables Scales of Measurement Exercises Solutions Simple Data Examination Tasks Theory Frequency Distribution of Discrete Variables Frequency Distribution of Continuous Variables Distribution Analysis with the Data Audit Node Concept of "SuperNodes" and Transforming a Variable to Normality Reclassifying Values Binning Continuous Data Exercises Solutions 259 Literature Multivariate Statistics Theory Scatterplot Scatterplot Matrix Correlation Correlation Matrix Exclusion of Spurious Correlations Contingency Tables Exercises Solutions 325 Literature Regression Models Introduction to Regression Models Motivating Examples Concept of the Modeling Process and Cross-Validation Simple Linear Regression Theory Building the Stream in SPSS Modeler Identification and Interpretation of the Model Parameters Assessment of the Goodness of Fit 362

4 ix Predicting Unknown Values Exercises Solutions Multiple Linear Regression Theory Building the Model in SPSS Modeler Final MLR Model and Its Goodness of Fit Prediction of Unknown Values Cross-Validation of the Model Boosting and Bagging (for Regression Models) Exercises Solutions Generalized Linear (Mixed) Model Theory Building a Model with the GLMM Node The Model Nugget Cross-Validation and Fitting a Quadric Regression Model Exercises Solutions The Auto Numeric Node Building a Stream with the Auto Numeric Node The Auto Numeric Model Nugget Exercises Solutions 500 Literature Factor Analysis Motivating Example General Theory of Factor Analysis Principal Component Analysis Theory Building a Model in SPSS Modeler Exercises Solutions Principal Factor Analysis Theory Building a Model Exercises Solutions 579 Literature Cluster Analysis Motivating Examples General Theory of Cluster Analysis 589

5 7.2.1 Exercises Solutions TwoStep Hierarchical Agglomerative Clustering Theoiy of Hierarchical Clustering Characteristics of the TwoStep Algorithm Building a Model in SPSS Modeler Exercises Solutions K-Means Partitioning Clustering Theory Building a Model in SPSS Modeler Exercises Solutions Auto Clustering Motivation and Implementation of the Auto Cluster Node Building a Model in SPSS Modeler Exercises Solutions Summary 710 Literature Classification Models Motivating Examples General Theory of Classification Models Process of Training and Using a Classification Model Classification Algorithms Classification vs. Clustering Making a Decision and the Decision Boundary Performance Measures of Classification Models The Analysis Node Exercises Solutions Logistic Regression Theory Building the Model in SPSS Modeler Optional: Model Types and Variable Interactions Final Model and Its Goodness of Fit Classification of Unknown Values Cross-Validation of the Model Exercises Solutions 758

6 xi 8.4 Linear Discriminate Classification Theory Building the Model with SPSS Modeler The Model Nugget and the Estimated Model Parameters Exercises Solutions Support Vector Machine Theory Building the Model with SPSS Modeler The Model Nugget Exercises Solutions Neuronal Networks Theory Building a Network with SPSS Modeler The Model Nugget Exercises Solutions k-nearest Neighbor Theory Building the Model with SPSS Modeler The Model Nugget Dimensional Reduction with PCA for Data Preprocessing Exercises Solutions Decision Trees Theory Building a Decision Tree with the C5.0 Node The Model Nugget Building a decision tree with the CHAID node Exercises Solutions The Auto Classifier Node Building a Stream with the Auto Classifier Node The Auto Classifier Model Nugget Exercises Solutions 974 Literature Using R with the Modeler Advantages of R with the Modeler Connecting with R Test the SPSS Modeler Connection to R Calculating New Variables in R 994

7 xii Contents 9.5 Model Building in R Exercises Solutions 1018 Literature Appendix Data Sets Used in This Book adult_income_data.txt beer.sav benchmark.xlsx car_simple.sav car_sales_modified.sav chess_endgame_data.txt customer_bank_data.csv diabetes_data_reduced.sav DRUGln.sav EEG_Sleep_Signals.csv employee_dataset_001 and employee_dataset_ England Payment Datasets Features_eeg_signals.csv gene_expression_leukemia.csv gene_expression_leukemia_short.csv gravity_constant_data.csv Housing.data.txt Iris.csv IT-projects.txt IT user satisfaction.sav longley.csv LPGA2009.csv Mtcars.csv nutrition_habites.sav optdigits_training.txt, optdigits_test.txt Orthodont.csv Ozone.csv pisa2012_math_q45.sav sales_list.sav ships.csv test_scores.sav Titanic.xlsx tree_credit.sav wine_data.txt WisconsinBreastCancerData.csv z_pm_customerl.sav 1057 Literature 1057