Device, system, and method for training machine learning algorithms to generate talkgroups based on in-call data

Abstract

A device, system and method for training machine-learning algorithms to generate talkgroups based on in-call data is provided. The device generates, via a machine-learning algorithm, an initial talkgroup based on an incident indication of an incident, the initial talkgroup comprising communication devices communicating via channels, the machine-learning algorithm initially trained to: generate talkgroups based on incident indications; and make changes to the talkgroups. The device receives in-call data such as communications in the initial talkgroup and/or call-transmission metadata, and generates, via the machine-learning algorithm, a suggested change to the initial talkgroup based on the in-call data. The device determines a machine-learning score of the initial talkgroup or the suggested change, indicative of a threshold-based positive or negative reinforcement of efficiency of the initial talkgroup or the suggested change, and implements a machine-learning feedback loop that provides the score to the machine-learning algorithm for further training thereof.