Upcoding detection in hospitals
Our system won a Government competition for detecting upcoding in hospitals – see here https://govtech.gov.pl/zdrowie/.
The system is based on Machine Learning exploration of patient records. The system finds anomalous patterns which are encoded as “detectors”. Each detector has a predefined format and provides intuitive explanations that can serve as a proof for audit purposes.
Context recommendation system
We took part in RecSys 2019 competition. We built a system for session based recommendations which currently is at the 1st place of this prestigious competition.
We specialize in event driven recommendations which adapts the suggestions to the last actions performed by the user.