- IDENTIFYING ACTIVITIES USING A HYBRID USER-ACTIVITY MODEL

摘要

In the user activity recognition technique, the user's actions are monitored while the user is using the computer. When these user actions are associated with user activities, the user activities are not initially specified and the tracked user actions constitute unmanaged data. Subsequently, the tracked user actions are aggregated into subsets (e.g., using clustering analysis) and user activity classifications (such as activity labels) for the subsets are provided by the user, Constitute managed data. Subsequently, when additional user behavior (associated with one or more initially not explicitly specified current user activities) is tracked, it may be associated with one or more classified subsets. For example, information about additional user behaviors may be mapped in real time (or near real time) to one or more subsets using a managed learning function. In this way, one or more current user activities can be identified.