DETECTION OF CREDIT CARD FRAUD IN REAL TIME USING SPARK ML

Assume you have access to a credit card. Your prior spending habits will be investigated. For example, how much money you spend, where it is spent, how frequently it is spent, and what you buy. If your current credit card transaction differs from your historical spending patterns, it is suspected of fraud; otherwise, it is considered as a genuine transaction, and fraud transactions are warned in the dashboard. The predictions are expected to be based on millions of transactions. As result, distributed frameworks that can expand as the number of transactions grows are used. Spark this system for real-time identification of credit card theft is built with Kafka and Cassandra. Preprocessing is accomplished through the use of Spark Machine Learning Pipeline Stages such as String Indexer and Vector.