Monitor Plus® Detection Technologies
False positive rate
Over 75% protection
Monitor Plus is the decisioning platform that incorporates best-of-breed proprietary expert judgement and Machine Learning technology and open Machine Learning technology through ONNX supporting the industry's best algorithms and respected data scientists.
The financial crime prevention models are based on white box technology that is fully open to the expert user and allows for the transparency, growth, flexibility and continuous evolution required by analysts, experts and regulators designed to adapt immediately to the emerging threats of a digital and global world.
"...our technology was designed from its foundation with the idea of being resilient over time, minimising dependency, to have the highest possible level of autonomy appropriate to the financial world we live in today. We did not believe at the beginning that this simple but powerful idea would be one of our greatest assets to our customers...".
State-of-the-art
Native Technology
Deep neural networks
- Brazilian Back Propagation.
- Up to five layers.
- Types of information as input nodes:
- Transaction information.
- DDS information.
- Expert model.
- Automatic and continuous re-training.
- Easy to parameterize and train.
Variables Engineering as Input to the
Machine Learning Technologies
Having historical information that is as rich and detailed as possible and with a high descriptive power to help algorithms recognise fraud is fundamental to train machine learning algorithms. The performance of these technologies is directly related to the input information provided to them.
Monitor Plus® information sources distinctive in the financial crime prevention solutions market:
Expert model
An expert model with more than 20 years of experience and maturity that contains all the learning obtained during the support of more than 400 financial institutions in 5 continents. This expert model provides Machine Learning technologies with accurate information about the characteristics of the transaction being analysed.
Dimensional data storage
Information cubes that perfectly describe and profile the credit/debit card, the cardholder, the device, the merchant, the ATM, etc. These are more than 3,500 variables that are automatically updated with each transaction performed by the card, cardholder, device, ATM, etc., act in conjunction with the expert model and form the second distinctive input for Machine Learning technologies.
Information flow and its enrichment to provide Machine Learning technologies with greater predictive power in their training and implementation in production:
Two Machine Learning Ecosystems
for better results
Native Ecosystem
ONNX RunTime Open Ecosystem
Deep neural networks
In-depth Knowledge of Customer Behavior
Naive Bayes
Monitor Plus® combines Machine Learning with a world-class expert system to provide 3:1 false positive rates and detection rates of over 75%. The platform provides a sandbox environment to test models before they go into production and uses Real Time with multiple Machine Learning models working as an assembly for best omni-channel results.
Easy and simple to use
Production Work Scheme
Incorporation of
ONNX RunTime
ONNX is the most powerful Deep Learning model exchange protocol on the market. It was created by Facebook and Microsoft in 2017 and is sponsored by all the major technology companies. This technology has been adopted by the market due to its high classification speed (it allows classifications in 3 milliseconds per transaction) and the vast ecosystem of algorithms that can be put into production.
