Open Conference Systems, ICQQMEAS2015

Font Size: 
Human Social Sensors: Delay-Driven Detection of Events in Social Media for Advertising Applications
Klimis Dalianis, Maria Tsirintani, Spyridon Binioris

Last modified: 2015-09-24

Abstract


Humans can be regarded as geographically distributed, multimodal sensors, who publish what they see and feel through blogs, forums, product reviews, and several available social networking sites. In this direction, Facebook or Twitter have short delays in reflecting what their users perceive, compared to other venues such as blogs and product reviews, which usually have much longer delays. And the question are: how good are the large social networks at near real-time sensing ? What are the lowest time limits of detecting a new event ? Towards this target, we devise efficient data collection and event recognition solutions over the most famous social networks (like Facebook, Twitter etc), so that content is rapidly extracted. These new capabilities are very useful not only e.g. for creating program guides for live broadcasting, but also for providing a better, more effective pricing mechanism for selling advertisement slots, typically by sensing instantaneous popularity of a segment. Preliminary experimental results over real data illustrate the promising performance of the proposed scheme

Full Text: PDF