Efficiently Discovering Hammock Paths from Induced Similarity Networks

ty networks are important abstractions in many information management applications such as recommender systems, corpora analysis, and medical informatics. For instance, in a recommender system, by inducing similarity networks between movies rated similarly by users, we can aim to find the global structure of connectivities underlying the data, and use the network to posit connections between given entities. We present an algorithmic framework to efficiently find paths in an induced similarity network without materializing the network in its entirety. Our framework introduces the notion of `hammock' paths which are generalizations of traditional paths in bipartite graphs. Given starting and ending objects of interest, it explores candidate objects for path following, and heuristics to admissibly estimate the potential for paths to lead to a desired destination. We present three diverse applications, modeled after the Netflix dataset, a broad subset of the PubMed corpus, and a data...

• Type: Journal
• Year: 2010
• Where: CORR
• Authors: M. Shahriar Hossain, Michael Narayan, Naren Ramakrishnan
• Keywords: CORR 2010, Education, Information Management Applications, Recommender System, Similarity Networks