News Stream Mining for Event Detection and Tracking and for Newswire-Timeliness Analysis

Day - Time: 10 October 2017, h.14:30
Place: Area della Ricerca CNR di Pisa - Room: C-29
  • Ida Mele (Università della Svizzera Italiana)

Raffaele Perego


ABSTRACT: Every day a huge volume of news are published online, therefore journalists and news analysts often require the support of automatic tools for coping with such a flood of information. This talk will focus on mining news published by several newswires on different platforms with the purpose of discovering and tracking events as well as of analyzing the temporal relationships among the news streams. The first part of the talk will present an approach based on Dynamic Topic Modeling for tackling the problems of detecting events and "chaining" them based on their evolution over time. In the second part, the attention will shift to cross-linking news streams for finding frequent temporal-publishing patterns and ranking the streams based on how timely the newswires report events on different platforms.

SHORT BIO: Ida Mele is a postdoctoral researcher in the Information Retrieval group at the Faculty of Informatics of the Università della Svizzera italiana (USI) in Lugano, Switzerland. Previously, she was a postdoctoral researcher at the Max Planck Institute for Informatics in Saarbruecken, Germany. She received her PhD from Sapienza University of Rome with a thesis on web usage mining for optimizing information retrieval systems and for improving webpage recommendation. Her research focuses on information retrieval, recommendation systems, web and social media mining.