- When: Jan 23 – Mar 15, 2018. This is a 3-credit, semester-length course that is scheduled at an accelerated pace of 8 weeks.
- Where: This is an on-line only course.
- Instructor: Prof. Marc Liberatore
- Enrollment not active yet.
This course provides an introduction to the principles and practice of fraud detection across a variety of problem domains such as money laundering, credit card fraud, telecommunications fraud, and computer and network intrusion. The key topics of this course include defining fraud in various domains; the interactions between fraud prevention and fraud detection; data collection and management; statistical tests and statistical power; methods for statistical fraud detection, including numerical validation, outlier detection, supervised and unsupervised classification methods, and application of Benford’s Law.
- Introduction and Course Overview: What is fraud? Why does it happen? Who is responsible for preventing and detecting it?
- Fraud Prevention: Risks that lead to fraud. Controls to prevent fraud.
- Symptoms of Fraud: Outliers, anomolies, implausibilities, and impossibilities.
- Auditing: Overview of auditing. Auditing plans, goals, possible outcomes. Computer-assisted auditing techniques.
- Dealing with Data: Collecting data. Cleaning and normalizing data. Data attributes and structure. Data integrity.
- Understanding Data: Computer analysis and techniques. Sorting, indexing, summarizing, and stratifying. Duplicate detection. Joining data tables. Pivot tables and cross tabulation.
- Symptoms of Known Fraud: Payroll fraud. Overpayments. Ven- dor/employee interaction. Kickbacks. Bid rigging.
- Testing for Outliers: Statistical tests. Statistical power. Randomization testing.
- Modeling Behavior: Supervised and unsupervised learning techniques for fraud detection.
- Other Test for Fraud: Frequently used values. Even amounts and rounding. Ratio/variance analysis. Min/maxes. Benford’s Law.
- Verifying Results and Improving Process: Sampling. Limits of computer- aided auditing. Preventative, detective, and corrective controls.