ControlCheck uses unsupervised machine learning algorithms to analyze large quantities of data across patients, locations, time, movement, and relationships. We find unsupervised machine learning advantageous over supervised machine learning.
With supervised machine learning, you must teach the AI what cases of diversion look like. This requires patterns of behavior from a large number of confirmed cases of diversion, which are then labeled as such to train the AI.
As users who divert medications become smarter in their tactics, unsupervised machine learning allows our software to continuously learn unusual and ever-changing patterns of behaviors. The AI continues to “evolve” it’s behavioral learning, allowing for the detection of patterns and outliers in data without the need for a large dataset of confirmed cases of diversion.