Applying Learning to Rank Techniques to Contextual Suggestions

The Text Retrieval Conference’s Contextual Suggestion Track investigates search techniques for complex information needs that are highly dependent on a context and user interests. The goal of the track is to evaluate systems that provide suggestions for activities to users in a specific location, taking into account their historical personal preferences. In this paper, we present our approach for the Contextual Suggestion Track 2014. We suggest to treat the problem of Contextual Suggestion as a Learning to Rank problem. As a source for travel suggestions we use data from four social networks: Yelp, Facebook, Foursquare and Google Places. For our study we train two ranking algorithms: Rank Net and Random Forest. In our experiments, we seek to answer the following research questions: Does the distance between the locations of training and testing contexts impact precision? Which data sources (i.e.,