NOTE - this example requires DSE Enterprise with Analytics enabled
This is an example of how to use DSE Analytics to detect fraudulent claims in a real time environment.
This example deals with a credit card processing application using Spark.MLlib to learn about fraudulent claims.
The application reads from a sample data set and then detect actual fraudulent claims from a separate data set containing simulated data.
- Start DSE in Analytics mode
- Run the included cql script resources/cql/create_schema.cql using cqlsh with the following command cqlsh -f '/resources/cql/create_schema.cql'
- Open the file src/main/resources/properties.txt
- set trainingDir= (located in src/main/resources/training)
- set testDir= (located in src/main/resources/test)
- set batchDuration and numFeatures to appropriate values. Defaults are respectively 1 and 3
To build the jar file run:
mvn clean package
To start the processor run
dse spark-submit <path to jar file> <path to properties.txt file>