Chapter 3. Building a Fraud Detection System
For fraud detection in financial systems, several commercial solutions are available on the market. However, Hadoop and its ecosystems of tools offer the opportunity to build a fraud detection system that can supplement existing fraud detection systems and lower their operational costs by downloading certain fraud detection tasks to Hadoop.
In this chapter, we have chosen Hadoop and Spark to build a simple fraud detection system. The following steps will be involved in building this solution:
- Selecting and cleansing the dataset.
- Designing the high-level architecture.
- Creating the fraud detection model.
- Putting the model to use.
You will learn how to build a machine learning model using Spark and Hadoop with the help of available datasets. You will use this model and Spark streaming to detect anomalies in real time that might be caused by fraudulent transactions.
In the upcoming sections of this chapter, we will cover how to build...
