Pairing of hearing devices with machine learning algorithm

A method for determining potential partner devices for a hearing device worn by a user, wherein the hearing device is adapted for establishing a data communication connection with other devices with a data communication interface and for receiving an audio data stream from the other devices via the data communication interface, the method comprising:

• generating context data from sensor data recorded by the hearing device;
• inputting the context data into a machine learning algorithm, which has been trained with historical context data to classify potential partner devices;
• detecting potential partner devices in an environment of the hearing device via the data communication interface and generating metadata for each detected potential partner device, wherein the metadata is additionally input into the machine learning algorithm that has been trained additionally with historical metadata to classify the potential partner devices;
• and
• outputting at least one classification by the machine learning algorithm, the classification classifying the potential partner devices and indicating whether the user expects a data communication connection with the potential partner devices.