Improving Tag-Based Recommendation by Topic Diversification Christian Wartena and martin Wibbels Noway, Brouwerijstraat 1, 7523 XC Enschede, The Netherlands Christian Wartena, Martin. Wibbels lenovay nl Abstract. Collaborative tagging has emerged as a mechanism to de- scribe items in large on-line collections. Tags are assigned by users to describe and find back items, but it is also tempting to describe the users in terms of the tags they assign or in terms of the tags of the items they are interested in. The tag-based profile thus obtained can be used to recommend new items If we recommend new items by computing their similarity to the user profile or to all items seen by the user, we run into the risk of recom- mending only neutral items that are a bit relevant for each topic a user is interested in. In order to increase user satisfaction many recommender systems not only optimize for accuracy but also for diversity. Often it assumed that there exists a trade-off between accuracy and diversity In this paper we introduce topic aware recommendation algorithms. Topic aware algorithms first detect different interests in the user profile and then generate recommendations for each of these interests. We study opic aware variants of three tag based recommendation algorithms and show that each of them gives better recommendations than their base variants, both in terms of precision and recall and in terms of diversity. 1 Introduction Collaborative tagging has emerged in the past decade as a mechanism to describe items in large collections available on-line. Tags are assigned by users to describe and find back previously viewed items. Thus a ternary relation between users tags and items is established. In this paper we will investigate some possibilities to construct a user profile based on tags, to identify distinct interests in this profile, and to recommend items relevant to those interests When we use a tag based user profile an item is recommended if it is relevant to II tags in the user profile. Similarly, if we use a collaborative filtering approach, we require the item to be similar to all items in the user profile. However, for a user who has some distinct interests, an item that fits the average of all his interests might be less accurate than an item that fits exactly one of his interests Thus we expect, at least in some cases, that recommendations will improve if we identify different interests in the user profile and take these into account for the recommendation of new items. If a recommendation strategy does so, re will call this strategy topic aware. In the following we will make a number P. Clough et al.(Eds ) ECIR 2011, LNCS 6611, pp. 43-54, 2011.Improving Tag-Based Recommendation by Topic Diversification Christian Wartena and Martin Wibbels Novay, Brouwerijstraat 1, 7523 XC Enschede, The Netherlands {Christian.Wartena,Martin.Wibbels}@novay.nl Abstract. Collaborative tagging has emerged as a mechanism to de￾scribe items in large on-line collections. Tags are assigned by users to describe and find back items, but it is also tempting to describe the users in terms of the tags they assign or in terms of the tags of the items they are interested in. The tag-based profile thus obtained can be used to recommend new items. If we recommend new items by computing their similarity to the user profile or to all items seen by the user, we run into the risk of recom￾mending only neutral items that are a bit relevant for each topic a user is interested in. In order to increase user satisfaction many recommender systems not only optimize for accuracy but also for diversity. Often it is assumed that there exists a trade-off between accuracy and diversity. In this paper we introduce topic aware recommendation algorithms. Topic aware algorithms first detect different interests in the user profile and then generate recommendations for each of these interests. We study topic aware variants of three tag based recommendation algorithms and show that each of them gives better recommendations than their base variants, both in terms of precision and recall and in terms of diversity. 1 Introduction Collaborative tagging has emerged in the past decade as a mechanism to describe items in large collections available on-line. Tags are assigned by users to describe and find back previously viewed items. Thus a ternary relation between users, tags and items is established. In this paper we will investigate some possibilities to construct a user profile based on tags, to identify distinct interests in this profile, and to recommend items relevant to those interests. When we use a tag based user profile an item is recommended if it is relevant to all tags in the user profile. Similarly, if we use a collaborative filtering approach, we require the item to be similar to all items in the user profile. However, for a user who has some distinct interests, an item that fits the average of all his interests might be less accurate than an item that fits exactly one of his interests. Thus we expect, at least in some cases, that recommendations will improve if we identify different interests in the user profile and take these into account for the recommendation of new items. If a recommendation strategy does so, we will call this strategy topic aware. In the following we will make a number P. Clough et al. (Eds.): ECIR 2011, LNCS 6611, pp. 43–54, 2011. c Springer-Verlag Berlin Heidelberg 2011