Machine Learning credit card dispute
Non-trivial filtering, aggregation, and decision logic applied to the models’ results
Nussknacker can be used where business teams want to define and run real-time decision algorithms, like in marketing, monitoring, Internet of Things, next best action, fraud detection, and more
Analysts should be able to configure the system themselves, deploy new rules and monitor their performance. Professional work also requires testing the set rules beforehand since blocking regular clients should be avoided at all costs
The Client needed both very high throughput and the ability to analyse events in time windows. After the competitive evaluation, the real-time marketing decisioning algorithms were implemented with Nussknacker.
Keeping up with the new conditions requires changes not only in terms of the approach to campaign creation, but also to the implementation process
In Nussknacker it is easy to set up obfuscated detection thresholds to prevent easy reverse engineering by fraudsters. Nussknacker uses the same expression language to formulate conditions for sensitivity thresholds, ranging from very simple to very complex.
Here are working demos with real-time data processing. For more details check the blog posts
When new fraud schemes emerge and new data sources need to be added, the implementation path is to add a new component to the interface.
The team is then quickly equipped with the right countermeasure.
Through the interface, changes to the business logic are deployed on the fly, without middlemen.
Your organisation can act and react without delay.
Your real-time marketing system can be designed for a faster innovation-deployment cycle
It’s a sound idea to enable business users to create or modify applications faster and at the same time relieve IT from mundane, repetitive tasks.