Machine Learning and Artificial Intelligence in Marketing and Sales: Essential Reference for Practitioners and Data Scientists

Description

Machine Learning and Artificial Intelligence in Marketing and Sales explores the ideas, and the statistical and mathematical concepts, behind Artificial Intelligence (AI) and machine learning models, as applied to marketing and sales, without getting lost in the details of mathematical derivations and computer programming. Bringing together the qualitative and the technological, and avoiding a simplistic broad overview, this book equips those in the field with methods to implement machine learning and AI models within their own organisations. Bridging the 'Domain Specialist - Data Scientist Gap'; (DS-DS Gap) is imperative to the success of this and chapters delve into this subject from a marketing practitioner and the data scientist perspective. Rather than a context-free introduction to AI and machine learning, data scientists implementing these methods for addressing marketing and sales problems will benefit most if they are exposed to how AI and machine learning have been applied specifically in the marketing and sales contexts. Marketing and sales practitioners who want to collaborate with data scientists can be much more effective when they expand their understanding across boundaries to include machine learning and AI. Cover Machine Learning and Artificial Intelligence in Marketing and Sales Praise for Machine Learning and Artificial Intelligence in Marketing and Sales Machine Learning and Artificial Intelligence in Marketing and Sales: Essential Reference for Practitioners and Data Scientists Copyright Dedication Table of Contents List of Figures, Tables and Illustrations Foreword Preface Acknowledgments Introduction 1. Introduction and Machine Learning Preliminaries: Training and Performance Assessment Chapter Outline 1. Training of Machine Learning Models 1.1 Regression and Classification Models 1.2 Cost Functions and Training of Machine Learning Models 1.3 Maximum Likelihood Estimation 1.4 Gradient-Based Learning 2. Performance Assessment for Regression and Classification Tasks 2.1 Performance Assessment for Regression Models 2.2 Performance Assessment for Classification 2.2.1 Percent Correctly Classified (PCC) and Hit Rate 2.2.2 Confusion Matrix 2.2.3 Receiver Operating Characteristics (ROC) Curve and Area under the Curve (AUC) 2.2.4 Cumulative Response Curve and Lift (Gains) Chart 2.2.5 Gini Coefficient Technical Detour 1 Technical Detour 2 2. Neural Networks in Marketing and Sales Chapter Outline 1. Introduction to Neural Networks 1.1 Early Evolution 1.2 The Neural Network Model 1.2.1 NN for Regression 1.2.2 NN for Classification 1.3 Cost Functions and Training of Neural Networks Using Backpropagation 1.4 Output Nodes 1.4.1 Linear Activation Function for Continuous Regression Outputs 1.4.2 Sigmoid Activation Function for Binary Outputs 1.4.3 Softmax Activation Function for Multiclass Outputs 2. Feature Importance Measurement and Visualization 2.1 Neural Interpretation Diagram (NID) 2.2 Profile Method for Sensitivity Analysis 2.3 Feature Importance Based on Connection Weights 2.4 Randomization Approach for Weight and Input Variable Significance 2.5 Feature Importance Based on Partial Derivatives 3. Applications of Neural Networks to Sales and Marketing 4. Case Studies Case Study 1: Churn Prediction Case Study 2: Rent Value Prediction Technical Detour 1 Technical Detour 2 Technical Detour 3 Technical Detour 4 Technical Detour 5 Technical Detour 6 Linear Activation Function for Continuous Regression Outputs Sigmoid Activations Function for Binary Outputs Softmax Activation Function for Multi-class Outputs 3. Overfitting and Regularization in Machine Learning Models Chapter Outline 1. Hyperparameters, Overfitting, Bias-variance Tradeoff, and Cross-validation 1.1 Hyperparameters 1.2 Overfitting 1.3 Bias-variance Tradeoff 1.4 Cross-validation 2. Regularization and Weight Decay 2.1 L2 Regularization 2.2 L1 Regularization 2.3 L1 and L2 Regularization as Constrained Optimization Problems 2.4 Regularization through Input Noise 2.5 Regularization through Early Stopping 2.6 Regularization through Sparse Representations 2.7 Regularization through Bagging and Other Ensemble Methods Technical Detour 1 Technical Detour 2 Technical Detour 3 Technical Detour 4 Weight Decay in L2 Regularization Weight Decay in L1 Regularization Technical Detour 5 Technical Detour 6 Technical Detour 7 Technical Detour 8 4. Support Vector Machines in Marketing and Sales Chapter Outline 1 Introduction to Support Vector Machines 1.1 Early Evolution 1.2 Nonlinear Classification Using SVM 2 Separating Hyperplanes 3 Role of Kernels in Machine Learning 3.1 Kernels as Measures of Similarity 3.2 Nonlinear Maps and Kernels 3.3 Kernel Trick 4 Optimal Separating Hyperplane 4.1 Margin between Two Classes 4.2 Maximal Margin Classification and Optimal Separating Hyperplane 5 Support Vector Classifier and SVM 6 Applications of SVM in Marketing and Sales 7 Case Studies Case Study 1: Consumer Choice Modeling Case Study 2: Rent Value vs Location Technical Detour 1 Technical Detour 2 Technical Detour 3 Technical Detour 4 Technical Detour 5 Technical Detour 6 Technical Detour 7 Technical Detour 8 Technical Detour 9 Technical Detour 10 Technical Detour 11 Illustration 3 Illustration 4 Illustration 5 5. Random Forest, Bagging, and Boosting of Decision Trees Chapter Outline 1. Early Evolution of Decision Trees: AID, THAID, CHAID 2. Classification and Regression Trees (CART) 2.1 Regression Trees 2.1.1 Greedy Algorithm 2.1.2 Cost Complexity Pruning 2.2 Classification Trees 3. Decision Trees and Segmentation 4. Bootstrapping, Bagging, and Boosting 4.1 Bootstrapping 4.2 Bagging 4.3 Boosting 5. Random Forest 6. Applications of Random Forests and Decision Trees in Marketing and Sales 7. Case Studies Case Study 1: Caravan Insurance Case Study 2: Wine Quality Technical detour 1: Technical detour 2: Technical detour 3: References Index