Machine Learning for Auditors: Automating Fraud Investigations Through Artificial Intelligence

Description

Intermediate user level Table of Contents About the Author About the Technical Reviewer Introduction Part I: Trusted Advisors Chapter 1: Three Lines of Defense AI, ML, and Auditing The Three Lines of Defense Model Risk Management Complexities The Three Lines Model Conclusion Chapter 2: Common Audit Challenges Data Literacy Manual Testing Data Sources Structured vs. Unstructured Data Citizen Developers Data Wrangling Data Bias Conclusion Chapter 3: Existing Solutions Substantive Testing CAATs Fit-for-Purpose Technologies Process Mining Continuous Auditing Conclusion Chapter 4: Data Analytics CRISP-DM Data Analytics Audit Applications Data Analytics vs. Data Science Conclusion Chapter 5: Analytics Structure and Environment Analytics Organization Structure Organization Climate The Role of Senior Leaders Conclusion Part II: Understanding Artificial Intelligence Chapter 6: Introduction to AI, Data Science, and Machine Learning A Self-Driving Car Components of an AI System CRISP-DM for Data Science Domain Knowledge Payment Fraud/Anomaly Detection Conclusion Chapter 7: Myths and Misconceptions Myth #1: You Need an Advanced Degree to Be a Data Scientist Myth #2: Correlation Implies Causation Myth #3: The Model Building Is the Most Critical Step Conclusion Chapter 8: Trust, but Verify What Is Trust, but Verify? Why Is It Important to Verify? Integrated Reporting Conclusion Chapter 9: Machine Learning Fundamentals Supervised Learning Classifiers Decision Trees Random Forests Support Vector Machines Logistic Regression Naive Bayes Deep Learning Confusion Matrix ROC Curves Regression Linear Regression Unsupervised Learning Clustering Algorithms k-means Clustering Hierarchical Clustering Silhouette Score Elbow Plot Dimensionality Reduction Curse of Dimensionality Principal Component Analysis Scree Plots Overfitting, Underfitting, and Feature Extraction Overfitting Underfitting Feature Extraction Ensemble Conclusion Chapter 10: Data Lakes Introduction to Data Lakes Tangible Value Role as Analytics Enabler Architectures Conclusion Chapter 11: Leveraging the Cloud Local Workstation Cloud Computing Amazon SageMaker Google Colab IBM Watson Conclusion Chapter 12: SCADA and Operational Technology Fourth Industrial Revolution SCADA Auditing Applying AI to SCADA Auditing Conclusion Part III: Storytelling Chapter 13: What Is Storytelling? Data Storytelling Common Pitfalls Misleading Graphs Anscombe’s Quartet Engaging the Audience Conclusion Chapter 14: Why Storytelling? Why Does It Work? General Guidelines of Good Storytelling General Dashboard Layout Conclusion Chapter 15: When to Use Storytelling? Use Stories to Inspire or Motivate an Action When Can We Use Storytelling? Less Is More Conclusion Chapter 16: Types of Visualizations Basic Visuals Advanced Visuals One-Hot Encoding Conclusion Chapter 17: Effective Stories Case Study: “The Best Stats You’ve Ever Seen” Case Study: “U.S. GUN KILLINGS IN 2018” Case Study: “Numbers of Different Magnitudes” Recap of Effective Storytelling Elements Conclusion Chapter 18: Storytelling Tools Technical Expertise Available Tools Qlik Power BI Tableau Mode Analytics Conclusion Chapter 19: Storytelling in Auditing Audit Use Cases Communication of Findings Support Recommendations Clarify Business Knowledge Conclusion Part IV: Implementation Recipes Chapter 20: How to Use the Recipes What Is a Recipe? Prerequisites Where Can You Find the Python Code? Implementation Considerations Conclusion Chapter 21: Fraud and Anomaly Detection The Dish: A Fraud and Anomaly Detection System Ingredients Instructions Step 1: Data Preparation Step 2: Exploratory Data Analysis Step 3: Apply Interquartile Range (IQR) Method Step 4: Perform Supervised Learning Step 5: Perform Unsupervised Learning Analysis Step 6: Review Exceptions with Additional Data Step 7: Re-evaluate the Models Variation and Serving Chapter 22: Access Management The Dish: ERP Access Management Audit Ingredients Instructions Step 1: Data Preparation Step 2: Exploratory Data Analysis Step 3: Scatter Plot of ERP Access Step 4: Review Exceptions with Additional Data Step 5: Reperform the Analysis Variation and Serving Chapter 23: Project Management The Dish: Project Portfolio Analysis Ingredients Instructions Step 1: Data Preparation Step 2: Exploratory Data Analysis Step 3: Perform Random Forest Classification Step 4: Review Feature Importance Variation and Serving Conclusion Chapter 24: Data Exploration The Dish: Understanding the Data Through Exploration Ingredients Instructions Step 1: Data Preparation Step 2: Exploratory Data Analysis Variation and Serving Conclusion Chapter 25: Vendor Duplicate Payments The Dish: Vendor Duplicate Payments Analysis Ingredients Instructions Step 1: Data Preparation Step 2: Perform K-NN Algorithm Step 3: Review Exceptions with Additional Data Variation and Serving Conclusion Chapter 26: CAATs 2.0 The Dish: CAATs Analysis Using ML Ingredients Instructions Step 1: Data Preparation Step 2: Exploratory Data Analysis Step 3: K-Means Clustering Variation and Serving Conclusion Chapter 27: Log Analysis The Dish: NLP Log Analysis Ingredients Instructions Step 1: Data Preparation Step 2: Exploratory Data Analysis Step 3: Perform Topic Modeling Step 4: Reperform the Analysis Variation and Serving Conclusion Chapter 28: Concluding Remarks Index