A simple graph based fraud detector.
Reads in an initial "batch" csv in order to load historical transactions.
Then reads in a "streaming" transactions csv that determines if each transaction is potentially fraudulent based on various features.
Writes the result of each feature application for every transaction to a corresponding txt file.
Adds the transaction after getting the results of each feature to the in memory graph.
- java 8
- maven 3
- clone repo
cd insight-paymomvn package
./run.sh
cd insight_testsuite./run_tests.sh