Machine Learning for Business Analytics

Galit Shmueli, PhD, is Chair Professor at National Tsing Hua University’s Institute of Service Science, College of Technology Management. She has designed and instructed business analytics courses since 2004 at University of Maryland, Statistics.com, The Indian School of Business, and National Tsing Hua University, Taiwan. Peter C. Bruce is the Founder and former President of the Institute for Statistics Education at Statistics.com. Peter Gedeck, PhD, is Senior Data Scientist at Collaborative Drug Discovery and Lecturer at the UVA School of Data Science. His speciality is the development of machine learning algorithms to predict biological and physicochemical properties of drug candidates. Nitin R. Patel, PhD, is cofounder and lead researcher at Cytel Inc. He was also a co-founder of Tata Consultancy Services. A Fellow of the American Statistical Association, Dr. Patel has served as a visiting professor at the Massachusetts Institute of Technology and at Harvard University. He is a Fellow of the Computer Society of India and was a professor at the Indian Institute of Management, Ahmedabad, for 15 years.

• Foreword by Gareth James xxiPreface to the Second Python Edition xxiiiAcknowledgments xxviiPart I PreliminariesChapter 1 Introduction 31.1 What Is Business Analytics? 31.2 What Is Machine Learning? 51.3 Machine Learning, AI, and Related Terms 51.4 Big Data 71.5 Data Science 81.6 Why Are There So Many Different Methods? 81.7 Terminology and Notation 91.8 Road Maps to This Book 12Chapter 2 Overview of the Machine Learning Process 172.1 Introduction 182.2 Core Ideas in Machine Learning 182.3 The Steps in a Machine Learning Project 222.4 Preliminary Steps 232.5 Predictive Power and Overfitting 372.6 Building a Predictive Model 432.7 Using Python for Machine Learning on a Local Machine 492.8 Automating Machine Learning Solutions 492.9 Ethical Practice in Machine Learning 54Part II Data Exploration and Dimension ReductionChapter 3 Data Visualization 613.1 Uses of Data Visualization 623.2 Data Examples 643.3 Basic Charts: Bar Charts, Line Charts, and Scatter Plots 663.4 Multidimensional Visualization 753.5 Specialized Visualizations 90Chapter 4 Dimension Reduction 1014.1 Introduction 1024.2 Curse of Dimensionality 1024.3 Practical Considerations 1034.4 Data Summaries 1034.5 Correlation Analysis 1084.6 Reducing the Number of Categories in Categorical Variables 1094.7 Converting a Categorical Variable to a Numerical Variable 1094.8 Principal Component Analysis 1114.9 Dimension Reduction Using Regression Models 1214.10 Dimension Reduction Using Classification and Regression Trees 121Part III Performance EvaluationChapter 5 Evaluating Predictive Performance 1295.1 Introduction 1305.2 Evaluating Predictive Performance 1315.3 Judging Classifier Performance 1375.4 Judging Ranking Performance 1505.5 Oversampling 156Part IV Prediction and Classification MethodsChapter 6 Multiple Linear Regression 1676.1 Introduction 1686.2 Explanatory vs. Predictive Modeling 1686.3 Estimating the Regression Equation and Prediction 1706.4 Variable Selection in Linear Regression 176Chapter 7 k-Nearest Neighbors (k-NN) 1937.1 The k-NN Classifier (Categorical Outcome) 1947.2 k-NN for a Numerical Outcome 2037.3 Advantages and Shortcomings of k-NN Algorithms 205Chapter 8 The Naive Bayes Classifier 2098.1 Introduction 2098.2 Applying the Full (Exact) Bayesian Classifier 2128.3 Solution: Naive Bayes 2138.4 Advantages and Shortcomings of the Naive Bayes Classifier 224Chapter 9 Classification and Regression Trees 2299.1 Introduction 2309.2 Classification Trees 2329.3 Evaluating the Performance of a Classification Tree 2419.4 Avoiding Overfitting 2469.5 Classification Rules from Trees 2529.6 Classification Trees for More Than Two Classes 2529.7 Regression Trees 2539.8 Advantages and Weaknesses of a Tree 2569.9 Improving Prediction: Random Forests and Boosted Trees 258Chapter 10 Logistic Regression 26710.1 Introduction 26810.2 The Logistic Regression Model 26910.3 Example: Acceptance of Personal Loan 27210.4 Evaluating Classification Performance 27710.5 Variable Selection 28010.6 Logistic Regression for Multi-Class Classification 28110.7 Example of Complete Analysis: Predicting Delayed Flights 285Chapter 11 Neural Nets 30111.1 Introduction 30211.2 Concept and Structure of a Neural Network 30211.3 Fitting a Network to Data 30311.4 Required User Input 31611.5 Exploring the Relationship Between Predictors and Outcome 31711.6 Deep Learning 31811.7 Advantages and Weaknesses of Neural Networks 329Chapter 12 Discriminant Analysis 33312.1 Introduction 33412.2 Distance of a Record from a Class 33612.3 Fisher's Linear Classification Functions 33712.4 Classification Performance of Discriminant Analysis 34112.5 Prior Probabilities 34212.6 Unequal Misclassification Costs 34212.7 Classifying More Than Two Classes 34412.8 Advantages and Weaknesses 347Chapter 13 Generating, Comparing, and Combining Multiple Models 35113.1 Ensembles 35213.2 Automated Machine Learning (AutoML) 35913.3 Explaining Model Predictions 36513.4 Summary 366Chapter 14 Experiments, Uplift Models, and Reinforcement Learning 37114.1 A/B Testing 37214.2 Uplift (Persuasion) Modeling 37714.3 Reinforcement Learning 38414.4 Summary 393Part V Mining Relationships Among RecordsChapter 15 Association Rules and Collaborative Filtering 39915.1 Association Rules 40015.2 Collaborative Filtering 41315.3 Summary 427Chapter 16 Cluster Analysis 43316.1 Introduction 43416.2 Measuring Distance Between Two Records 43716.3 Measuring Distance Between Two Clusters 44316.4 Hierarchical (Agglomerative) Clustering 44516.5 Non-Hierarchical Clustering: The k-Means Algorithm 453Part VI Forecasting Time SeriesChapter 17 Handling Time Series 46317.1 Introduction 46417.2 Descriptive vs. Predictive Modeling 46517.3 Popular Forecasting Methods in Business 46517.4 Time Series Components 46617.5 Data Partitioning and Performance Evaluation 470Chapter 18 Regression-Based Forecasting 47718.1 A Model with Trend 47818.2 A Model with Seasonality 48418.3 A Model with Trend and Seasonality 48618.4 Autocorrelation and ARIMA Models 488Chapter 19 Smoothing and Deep Learning Methods for Forecasting 50919.1 Smoothing Methods: Introduction 51019.2 Moving Average 51019.3 Simple Exponential Smoothing 51519.4 Advanced Exponential Smoothing 51819.5 Deep Learning for Forecasting 521Part VII Data AnalyticsChapter 20 Social Network Analytics 53720.1 Introduction 53820.2 Directed vs. Undirected Networks 53820.3 Visualizing and Analyzing Networks 53920.4 Social Data Metrics and Taxonomy 54420.5 Using Network Metrics in Prediction and Classification 55020.6 Business Uses of Social Network Analysis 55620.7 Summary 557Chapter 21 Text Mining 56121.1 Introduction 56221.2 The Tabular Representation of Text 56221.3 Bag-of-Words vs. Meaning Extraction at Document Level 56321.4 Preprocessing the Text 56421.5 Implementing Machine Learning Methods 57321.6 Example: Online Discussions on Autos and Electronics 57321.7 Deep Learning Approaches 57721.8 Example: Sentiment Analysis of Movie Reviews 57821.9 Summary 581Chapter 22 Responsible Data Science 58722.1 Introduction 58822.2 Unintentional Harm 58922.3 Legal Considerations 59122.4 Principles of Responsible Data Science 59222.5 A Responsible Data Science Framework 59522.6 Documentation Tools 59922.7 Example: Applying the RDS Framework to the COMPAS Example 60322.8 Summary 613Chapter 23 Generative AI 61723.1 The Transformative Power of Generative AI 61723.2 What is Generative AI? 61923.3 Data and Infrastructure Requirements 62123.4 Adapting Models for Specific Purposes 62323.5 Prompt Engineering 62423.6 Uses of Generative AI 62523.7 Caveats and Concerns 62923.8 Summary 631Part VIII CasesChapter 24 Cases 63924.1 Charles Book Club 63924.2 German Credit 64624.3 Tayko Software Cataloger 65124.4 Political Persuasion 65524.5 Taxi Cancellations 65924.6 Segmenting Consumers of Bath Soap 66124.7 Direct-Mail Fundraising 66524.8 Catalog Cross-Selling 66824.9 Time-Series Case: Forecasting Public Transportation Demand 67024.10 Loan Approval 672References 675Index 677