Just Enough R!: An Interactive Approach to Machine Learning and Analytics

Cover
Half Title
Title Page
Copyright Page
Table of Contents
Preface
Acknowledgment
Author
Chapter 1 Introduction to Machine Learning
	1.1 Machine Learning, Statistical Analysis, and Data Science
	1.2 Machine Learning: A First Example
		1.2.1 Attribute-Value format
		1.2.2 A Decision Tree for Diagnosing Illness
	1.3 Machine Learning Strategies
		1.3.1 Classicisation
		1.3.2 Estimation
		1.3.3 Prediction
		1.3.4 Unsupervised Clustering
		1.3.5 Market Basket Analysis
	1.4 Evaluating Performance
		1.4.1 Evaluating Supervised Models
		1.4.2 Two-Class Error Analysis
		1.4.3 Evaluating Numeric Output
		1.4.4 Comparing Models by Measuring Lift
		1.4.5 Unsupervised Model Evaluation
	1.5 Ethical Issues
	1.6 Chapter Summary
	1.7 Key Terms
	Exercises
Chapter 2 Introduction to R
	2.1 Introducing R And RStudio
		2.1.1 Features of R
		2.1.2 Installing R
		2.1.3 Installing RStudio
	2.2 Navigating RStudio
		2.2.1 The Console
		2.2.2 The Source Panel
		2.2.3 The Global Environment
		2.2.4 Packages
	2.3 Where’s The Data?
	2.4 Obtaining Help and Additional Information
	2.5 Summary
	Exercises
Chapter 3 Data Structures and Manipulation
	3.1 Data Type
		3.1.1 Character Data and Factors
	3.2 Single-Mode Data Structures
		3.2.1 Vectors
		3.2.2 Matrices and Arrays
	3.3 Multimode Data Structures
		3.3.1 Lists
		3.3.2 Data Frames
	3.4 Writing Your Own Functions
		3.4.1 Writing a Simple Function
		3.4.2 Conditional Statements
		3.4.3 Iteration
		3.4.4 Recursive Programming
	3.5 Summary
	3.6 Key Terms
	Exercises
Chapter 4 Preparing the Data
	4.1 A Process Model for Knowledge Discovery
	4.2 Creating A Target Dataset
		4.2.1 Interfacing R with the Relational Model
		4.2.2 Additional Sources for Target Data
	4.3 Data Preprocessing
		4.3.1 Noisy Data
		4.3.2 Preprocessing With R
		4.3.3 Detecting Outliers
		4.3.4 Missing Data
	4.4 Data Transformation
		4.4.1 Data Normalization
		4.4.2 Data Type Conversion
		4.4.3 Attribute and Instance Selection
		4.4.4 Creating Training and Test Set Data
		4.4.5 Cross Validation and Bootstrapping
		4.4.6 Large-Sized Data
	4.5 Chapter Summary
	4.6 Key Terms
	Exercises
Chapter 5 Supervised Statistical Techniques
	5.1 Simple Linear Regression
	5.2 Multiple Linear Regression
		5.2.1 Multiple Linear Regression: An Example
		5.2.2 Evaluating Numeric Output
		5.2.3 Training/Test Set Evaluation
		5.2.4 Using Cross Validation
		5.2.5 Linear Regression with Categorical Data
	5.3 Logistic Regression
		5.3.1 Transforming the Linear Regression Model
		5.3.2 The Logistic Regression Model
		5.3.3 Logistic Regression with R
		5.3.4 Creating a Confusion Matrix
		5.3.5 Receiver Operating Characteristics (ROC) Curves
		5.3.6 The Area under an ROC Curve
	5.4 Naïve Bayes Classifier
		5.4.1 Bayes Classifier: An Example
		5.4.2 Zero-Valued Attribute Counts
		5.4.3 Missing Data
		5.4.4 Numeric Data
		5.4.5 Experimenting With Naïve Bayes
	5.5 Chapter Summary
	5.6 Key Terms
	Exercises
Chapter 6 Tree-Based Methods
	6.1 A Decision Tree Algorithm
		6.1.1 An Algorithm for Building Decision Trees
		6.1.2 C4.5 Attribute Selection
		6.1.3 Other Methods for Building Decision Trees
	6.2 Building Decision Trees: C5.0
		6.2.1 A Decision Tree for Credit Card Promotions
		6.2.2 Data for Simulating Customer Churn
		6.2.3 Predicting Customer Churn with C5.0
	6.3 Building Decision Trees: Rpart
		6.3.1 An Rpart Decision Tree for Credit Card Promotions
		6.3.2 Train and Test Rpart: Churn Data
		6.3.3 Cross Validation Rpart: Churn Data
	6.4 Building Decision Trees: J48
	6.5 Ensemble Techniques for Improving Performance
		6.5.1 Bagging
		6.5.2 Boosting
		6.5.3 Boosting: An Example with C5.0
		6.5.4 Random Forests
	6.6 Regression Trees
	6.7 Chapter Summary
	6.8 Key Terms
	Exercises
Chapter 7 Rule-Based Techniques
	7.1 From Trees to Rules
		7.1.1 The Spam Email Dataset
		7.1.2 Spam Email Classification: C5.0
	7.2 A Basic Covering Rule Algorithm
		7.2.1 Generating Covering Rules With JRip
	7.3 Generating Association Rules
		7.3.1 Confidence and Support
		7.3.2 Mining Association Rules: An Example
		7.3.3 General Considerations
		7.3.4 Rweka’s Apriori Function
	7.4 Shake, Rattle, and Roll
	7.5 Chapter Summary
	7.6 Key Terms
	Exercises
Chapter 8 Neural Networks
	8.1 Feed-Forward Neural Networks
		8.1.1 Neural Network Input Format
		8.1.2 Neural Network Output Format
		8.1.3 The Sigmoid Evaluation Function
	8.2 Neural Network Training: A Conceptual View
		8.2.1 Supervised Learning with Feed-Forward Networks
		8.2.2 Unsupervised Clustering With Self-Organizing Maps
	8.3 Neural Network Explanation
	8.4 General Considerations
		8.4.1 Strengths
		8.4.2 Weaknesses
	8.5 Neural Network Training: A Detailed View
		8.5.1 The Backpropagation Algorithm: An Example
		8.5.2 Kohonen Self-Organizing Maps: An Example
	8.6 Building Neural Networks with R
		8.6.1 The Exclusive-OR Function
		8.6.2 Modeling Exclusive-OR With MLP: Numeric Output
		8.6.3 Modeling Exclusive-OR With MLP: Categorical Output
		8.6.4 Modeling Exclusive-OR With Neuralnet: Numeric Output
		8.6.5 Modeling Exclusive-OR With Neuralnet: Categorical Output
		8.6.6 Classifying Satellite Image Data
		8.6.7 Testing For Diabetes
	8.7 Neural Net Clustering For Attribute Evaluation
	8.8 Times Series Analysis
		8.8.1 Stock Market Analytics
		8.8.2 Time Series Analysis: An Example
		8.8.3 The Target Data
		8.8.4 Modeling the Time Series
		8.8.5 General Considerations
	8.9 Chapter Summary
	8.10 Key Terms
	Exercises
Chapter 9 Formal Evaluation Techniques
	9.1 What Should Be Evaluated?
	9.2 Tools for Evaluation
		9.2.1 Single-Valued Summary Statistics
		9.2.2 The Normal Distribution
		9.2.3 Normal Distributions and Sample Means
		9.2.4 A Classical Model for Hypothesis Testing
	9.3 Computing Test Set Confidence Intervals
	9.4 Comparing Supervised Models
		9.4.1 Comparing the Performance of Two Models
		9.4.2 Comparing the Performance of Two or More Models
	9.5 Confidence Intervals for Numeric Output
	9.6 Chapter Summary
	9.7 Key Terms
	Exercises
Chapter 10 Support Vector Machines
	10.1 Linearly Separable Classes
	10.2 The Nonlinear Case
	10.3 Experimenting With Linearly Separable Data
	10.4 Microarray Data Mining
		10.4.1 DNA and Gene Expression
		10.4.2 Preprocessing Microarray Data: Attribute Selection
		10.4.3 Microarray Data Mining: Issues
	10.5 A Microarray Application
		10.5.1 Establishing a Benchmark
		10.5.2 Attribute Elimination
	10.6 Chapter Summary
	10.7 Key Terms
	Exercises
Chapter 11 Unsupervised Clustering Techniques
	11.1 The K-Means Algorithm
		11.1.1 An Example Using K-Means
		11.1.2 General Considerations
	11.2 Agglomerative Clustering
		11.2.1 Agglomerative Clustering: An Example
		11.2.2 General Considerations
	11.3 Conceptual Clustering
		11.3.1 Measuring Category Utility
		11.3.2 Conceptual Clustering: An Example
		11.3.3 General Considerations
	11.4 Expectation Maximization
	11.5 Unsupervised Clustering With R
		11.5.1 Supervised Learning for Cluster Evaluation
		11.5.2 Unsupervised Clustering For Attribute Evaluation
		11.5.3 Agglomerative Cluster: A Simple Example
		11.5.4 Agglomerative Clustering of Gamma-Ray Burst data
		11.5.5 Agglomerative Clustering Of Cardiology Patient Data
		11.5.6 Agglomerative Clustering Of Credit Screening Data
	11.6 Chapter Summary
	11.7 Key Terms
	Exercises
Chapter 12 A Case Study in Predicting Treatment Outcome
	12.1 Goal Identification
	12.2 A Measure of Treatment Success
	12.3 Target Data Creation
	12.4 Data Preprocessing
	12.5 Data Transformation
	12.6 Data Mining
		12.6.1 Two-Class Experiments
	12.7 Interpretation and Evaluation
		12.7.1 Should Patients Torso Rotate?
	12.8 Taking Action
	12.9 Chapter Summary
Bibliography
Appendix A: Supplementary Materials and More Datasets
Appendix B: Statistics for Performance Evaluation
Subject Index
Index of R Functions
Script Index