Fraud Detection in R

The Association of Certified Fraud Examiners estimates that fraud costs companies $3.7 trillion each year and accounts for five percent of the average company's annual revenues. In the future, fraud attempts are expected to increase. It is imperative that all industries are able to detect fraud. This course will show you how historical data can help detect fraud. We will use robust statistics and digit analysis to detect suspicious patterns that could indicate fraud. The main issues in developing a supervised tool for detecting fraud are the skewness and imbalance of the data as well as the misclassification costs. These problems can be addressed using techniques that use real and artificial data from a variety fraud applications.