Planning-context aware mobile recommendations

Abstract

In the realm of mobile applications a significant effort has been made to develop recommender systems that customize results based off of one’s current location and more recently even their inferred current activity. While this aspect of context has been shown to be quite successful, we suggest anticipating what they are currently planning for the future may help further improve the relevancy of the results as well. This work examines this problem as one of trying to predict the user’s planning context, defined as what activities are currently being planned and how far in the future the event they are planning is going to be. An empirical analysis is made of the predictability of planning context and a discussion of the potential implications of this for mobile context aware recommenders.