A system for suggesting local activities based on inferred user and location attributes (e.g. familiarity with an area, time of day, day of week), as well as explicitly stated attributes (e.g. group size + composition, such as alone, couple, with friends, with kids) is presented. The system is based on an ontology of activities including highlevel intents, moods etc. afforded by a particular locality. These activities are then mapped to contextual factors by expert editors who assign a value to each activity and contextual factor intersection, indicating the degree to which an activity is suited for a particular context. The system then suggests activities based on the inferred and explicit userstated contextual factors using the mapping. Advantages of the system include generation of expert suggestions or recommendations that are similar to what people provide one another, which encourages discovery by highlighting locally typical activities.
Varady, Gergely; Seefeld, Bernhard; and Riegelsberger, Jens, "SUGGESTING LOCAL ACTIVITIES BY INFERRED CONTEXT", Technical Disclosure Commons, (August 24, 2016)