An Inverse Collaborative Filtering Approach for Cold-Start Problem in Web Service Recommendation
Lianyong Qi,Wan-Chun Dou,Xuyun Zhang
DOI: https://doi.org/10.1145/3014812.3014860
2017-01-01
Abstract:Due to the increasing volume and variety of web services in different service communities, users are apt to find their interested web services through various recommendation techniques, e.g., Collaborative Filtering (i.e., CF) recommendation. In CF (e.g., user-based CF, item-based CF or hybrid CF) recommendation, "the similar friends of target user" or "the similar services of target services (i.e., the services preferred by target user)" are determined first; afterwards, "the services preferred by similar friends" or "the similar services of target services" are recommended to the target user. However, due to the inherent data sparsity in service recommendation, cold-start problem is inevitable when the target user has no similar friends and the target services have no similar services. While present CF recommendation approaches cannot deal with this cold-start problem very well. In view of this shortcoming, in this paper, a novel inverse CF approach named Inverse_CF_Rec is introduced to help alleviate the cold-start problem in service recommendation. Concretely, in Inverse_CF_Rec, we first look for the target user's enemy (i.e., antonym of "friend"), and then determine the target user's "possible friends" based on Social Balance Theory (e.g., "enemy's enemy is a friend" rule). Afterwards, "the services preferred by "possible friends" of target user" or "the services disliked by enemies of target user" are recommended to the target user, so as to alleviate the cold-start problem. Finally, through a set of simulation experiments deployed on well-known MovieLens-1M dataset, we validate the feasibility of our proposal in terms of recommendation accuracy, recall and efficiency.