Fraud Detection For A Major Insurance Company
The insurer was using Big-Data and had already an algorithm trained over 10+ years of data. The challenge was therefore to fully understand their existing algorithm and whole their whole ETL process was structured.
How London Data Consulting Helped
The London Data Consulting have improved both their database structure and ETL process so quicker, more optimal and comprehensive calculations could be performed.
The use of Kafka and Spark MLLib allowed us to perform quicker and more optimal predictions.
We then improved the existing algorithm with the use of deep learning and convolutional neural networks (CNN). We eventually improved the prediction accuracy by 30%.