TrustWalker: A Random Walk Model for Combining Trust-based and Item-based Recommendation
Jamali, M., & Ester, M.
ACM SIGKDD '09
http://www.cs.sfu.ca/~ester/papers/KDD-2009-TrustWalker.final.pdf
ABSTRACT
Collaborative filtering is the most popular approach to build recommender systems and has been successfully employed in many applications. However, it cannot make recommendations for so-called cold start users that have rated only a very small number of items. In addition, these methods do not know how confident they are in their recommendations. Trust-based recommendation methods assume the additional knowledge of a trust network among users and can better deal with cold start users, since users only need to be simply connected to the trust network. On the other hand, the sparsity of the user item ratings forces the trust-based approach to consider ratings of indirect neighbors that are only weakly trusted, which may decrease its precision. In order to find a good trade-off, we propose a random walk model combining the trust-based and the collaborative filtering approach for recommendation. The random walk model allows us to define and to measure the confidence of a recommendation. We performed an evaluation on the Epinions dataset and compared our model with existing trust-based and collaborative filtering methods.
1. INTRODUCTION
With the rapidly growing amount of information available on the WWW, it becomes necessary to have tools to help users to select the relevant part of online information. To satisfy this need, recommender systems have emerged, e.g. there are popular recommenders for movies1, books2, music3, etc.
Typically in a recommender system, we have a set of users and a set of items. Each user u rates a set of items by some values. The recommender has the task to predict the rating for user u on a non-rated item i or to generally recommend some items for the given user u based on the ratings that already exist. Collaborative Filtering [6] methods make recommendations based on the ratings of item i by a set of users whose rating profiles are most similar to that of user u. With the advent of online social networks, the trust-based approach to recommendation has emerged. This approach assumes a trust network among users and makes recommendations based on the ratings of the users that are directly or indirectly trusted by u.
Collaborative Filtering is most effective when users have expressed enough ratings to have common ratings with other users, but it performs poorly for so-called cold start user. Cold start users are new users who have expressed only a few ratings. Using similarity based approaches, it is unlikely to find similar users since the cold start users only have a few ratings. Trust-based recommenders, however, can make recommendations as long as a new user is connected to a large enough component of the trust network.
Using a trust network therefore improves the coverage of recommendations. However, when we go far from the source user u in the trust network, the trust between these users and the source user will become fairly weak and their ratings will be noisy and unreliable. Therefore, we have to use the ratings expressed by users in the neighborhood close to the user u. But, in this case the probability of finding a rating expressed on the item will be very low and we will not be able to compute a prediction.
In order to consider enough ratings without suffering from noisy data, we propose a random walk method (TrustWalker) which combines trust-based and item-based recommendation. TrustWalker considers not only ratings of the target item, but also those of similar items. The probability of using the rating of a similar item instead of a rating for the target item increases with increasing length of the walk. Our framework contains both trust-based and item-based collaborative filtering recommendations as special cases. Most traditional recommender systems do not provide confidence in their predictions. The random walk model allows us to compute the confidence in our predictions. Also our system is able to explain and justify its results. The rest of the paper is organized as follows: Section 2 describes the problem definition. Related works are discussed in section 3. We discuss the details of our proposed model in section 4. In section 5 we introduce some desirable properties of our model. The experimental results and comparison with existing methods are discussed in section 6. Finally we conclude the paper in section 7 and introduce some directions for future research.
TrustWalker is the random walk model which is the combination of C.F. and Trust-based method.
Ratings expressed by strongly trusted friends on similar item is more reliable than the ratings expressed by weakly trusted for neighbors on the target item.
There is a trade-off between precision and converge.
References
- none