Scalable, continuous tracking of tag co-occurrences between short sets using (almost) disjoint tag partitions

In this work we consider the continuous computation of set correlations over a stream of set-valued attributes, such as Tweets and their hashtags, social annotations of blog posts obtained through RSS, or updates to set-valued attributes of databases. In order to compute tag correlations in a distributed fashion, all necessary information has to be present at the computing node(s). Our approach makes use of a partitioning scheme based on set covers for efficient and replication-lean information flow. We report on the results of a preliminary performance evaluation using Tweets obtained through Twitter's streaming API.