Fraud Outlier Detection
Anticipate malicious actions
Decrease financial losses
Restore customer trust
What is Fraud Detection?
Detecting anomalies consists in identifying atypical items by cross-referencing your data in order to discover potentially fraudulent behaviors.
A system of outlier detection allows you to reliably rank fraud suspicions in order to be more efficient in their processing.
How Saagie helps you implementing a system of fraud detection?
This detection is made possible by centralizing a wide variety of data sources in the Saagie data lake, including:
- Banking transactions
- Intelligence shared by financial institutions
- Social media activities
- Processing of natural language
If a balanced history of referenced fraud is available, Saagie allows you to compare machine learning algorithms in the language used by your actuaries and data scientists. Even in the absence of any history it is possible to use other families of techniques to reveal suspicious behaviors.
The strongest suspicions are then submitted to further verifications combining both human expertise and automated processes. Once this verification is done, you can use Saagie to create a feedback mechanism thanks to which the advisor can confirm or deny the fraud, hence strengthening the detection algorithm on a daily basis. This is an online machine learning algorithm (self-learning).
Fraud Detection with Matmut
Thanks to Saagie, we succeeded in improving our fraud detection system by cross-referencing more data sources (compared to what we were doing before).
As a result, we helped our advisors to save a lot of time by reducing the number of fraud suspicions later proven wrong (false positives); they were then able to focus on the most critical cases of fraud.