Patent attributes
A computer-implemented method according to one embodiment includes identifying one or more historical instances of device usage; determining historical contextual data for each of the one or more historical instances of the device usage; creating an event in association with the one or more historical instances of the device usage and the historical contextual data; training a classifier to identify the event by inputting the historical contextual data and an associated event identifier into a machine learning algorithm; identifying, using the trained classifier, a real-time occurrence of the event, utilizing real-time contextual data and the historical contextual data, where the trained classifier analyzes the real-time contextual data to identify the real-time occurrence of the event; and in response to identifying the real-time occurrence of the event, adjusting one or more aspects of a user interface of the device based on the one or more historical instances of the device usage.