http://www.hpl.hp.com/research/scl/papers/bundles/bundles.pdf Clusters in social networks can help design customized bundles of products for consumers. Identifying consumer preferences is a key challenge in customizing electronic commerce sites to individual users. The increasing availability of online social networks provides one approach to this problem: people linked in these networks often share preferences, allowing inference of interest in products based on knowledge of a consumer's network neighbors and their interests. This paper evaluates the benefits of inference from online social networks in two contexts: a random graph model and a web site allowing people to both express preferences and form distinct social and preference links. We determine conditions on network topology and preference correlations leading to extended clusters of people with similar interests. Knowledge of when such clusters occur improves the usefulness of social network-based inference for identifying products likely to interest consumers based on information from a few people in the network. Such estimates could help sellers design customized bundles of products and improve combinatorial auctions for complementary products. To appear in Electronic Commerce Research and Applications special issue on Social Networks. Tad Hogg 2009-04-27 21:09:00 social networks electronic commerce essembly EC http://www.hpl.hp.com/research/scl/papers/sncf/ User-generated social networking links can be as predictive as algorithmically identified "neighbors" in recommender systems. collaborative filtering social networks prediction recommender systems essembly This paper demonstrates that "social network collaborative filtering" (SNCF), wherein user-selected like-minded alters are used to make predictions, can rival traditional user-to-user collaborative filtering (CF) in predictive accuracy. Using a unique data set from an online community where users rated items and also created social networking links specifically intended to represent like-minded allies, we use SNCF and traditional CF to predict ratings by networked users. We find that SNCF using generic "friend" alters is moderately worse than the better CF techniques, but outperforms benchmarks such as by-item or by-user average rating; generic friends often are not like-minded. However, SNCF using "ally" alters is competitive with CF. These results are significant because SNCF is tremendously more computationally efficient than traditional user-user CF and may be implemented in large-scale web commerce and social networking communities. It is notoriously difficult to distinguish the contributions of social influence (where allies influence users) and social selection (where users are simply effective at selecting like-minded people as their allies). Nonetheless, comparing similarity over time, we do show no evidence of strong social influence among allies or friends. Rong Zheng, Dennis M. Wilkinson and Foster Provost 2008-10-06 12:00:00 Tad Hogg and Gabor Szabo 2008-03-25 0:00:00 Web sites where users create and rate content as well as form networks with other users display long-tailed distributions in many aspects of behavior. Using behavior on one such community site, Essembly, we propose and evaluate plausible mechanisms to explain these behaviors. Unlike purely descriptive models, these mechanisms rely on user behaviors based on information available locally to each user. For Essembly, we find the long-tails arise from large differences among user activity rates and qualities of the rated content, as well as the extensive variability in the time users devote to the site. We show that the models not only explain overall behavior but also allow estimating the quality of content from their early behaviors. http://www.hpl.hp.com/research/idl/papers/diversity/ Diversity among users and the content they create in the Essembly web site. social networks essembly Tad Hogg, Gabor Szabo, Dennis M. Wilkinson, and Michael J. Brzozowski 2008-01-12 0:00:00 Traditionally, online social network sites like Facebook and MySpace allow people to form links to "friends" but do little to qualify the semantic meaning of the friendship. As a result, many users "collect" friends on these sites, conflating "acquaintances" with "friends". Essembly, a "fiercely non-partisan social network", on the other hand, lets its users enrich the meaning of their relations to others by explicitly labeling them "friends", "allies", or "nemeses". Essembly then allows its members to post resolves reflecting controversial opinions on political issues. As a defining activity on the site, members can vote on these resolves on a four-point scale ranging from complete agreement to full opposition. We examined how the uncommon link semantics affects users in casting their votes. In particular, Essembly prominently highlights the activities of users' acquaintances, and the question arises if this makes them more likely to participate, and if so, how this information affects votes. It is widely assumed that social networks enhance, if not drive, the popularity of online services; what does an additional layer of link classification add to them? Papers appeared at CHI 2008 and AAAI Spring Symposium on Social Information Processing 2008. http://www.hpl.hp.com/research/idl/papers/essembly Examines the usefulness of distinguishing between friends and similar/dissimilar users in propagating new content in an online social network, and suggests resulting design implications for social content aggregation services and recommender systems. social networks voting essembly influence CHI