Abstract:”Social search is having a flourishing success for its effectiveness in retrieving high quality information useful to achieve complex search goals. Surprisingly, the potential of the social paradigm at the basis of collaborative tagging in satisfying complex search intents has been unexplored so far. We propose an extended model of folksonomies that allows to compose and tag complex user-defined relations between items, users, and tags. We show that this model offers several means to fulfill complex search tasks that are hard to be achieved by other existing services. Furthermore, we support the validity of our approach through a data-driven analysis on Flickr photosets and we present an online portal that provides this new user experience. “
Our paper “On the dynamics of human proximity for data diffusion in ad-hoc networks” accepted in the Ad Hoc Networks journal is available online: doi:10.1016/j.adhoc.2011.06.003
Abstract: We report on a data-driven investigation aimed at understanding the dynamics of message spreading in a real-world dynamical network of human proximity. We use data collected by means of a proximity-sensing network of wearable sensors that we deployed at three different social gatherings, simultaneously involving several hundred individuals. We simulate a message spreading process over the recorded proximity network, focusing on both the topological and the temporal properties. We show that by using an appropriate technique to deal with the temporal heterogeneity of proximity events, a universal statistical pattern emerges for the delivery times of messages, robust across all the data sets. Our results are useful to set constraints for generic processes of data dissemination, as well as to validate established models of human mobility and proximity that are frequently used to simulate realistic behaviors.
Abstract: “Categorization of web-search queries in semantically coherent topics is a crucial task to understand the interest trends of search engine users and, therefore, to provide more intelligent personalization services. Query clustering usually relies on lexical and clickthrough data, while the information originating from the user actions in submitting their queries is currently neglected. In particular, the intent that drives users to submit their requests is an important element for meaningful aggregation of queries. We propose a new intentcentric notion of topical query clusters and we define a query clustering technique that differs from existing algorithms in both methodology and nature of the resulting clusters. Our method extracts topics from the query log by merging mis- sions, i.e., activity fragments that express a coherent user intent, on the basis of their topical affinity. Our approach works in a bottom-up way, without any a-priori knowledge of topical categorization, and produces good quality topics compared to state-of-the-art clustering techniques. It can also summarize topically-coherent missions that occur far away from each other, thus enabling a more compact user profiling on a topical basis. Furthermore, such a topical user profiling discriminates the stream of activity of a particular user from the activity of others, with a potential to predict future user search activity. “