Detecting Multiple Facets of an Event using Graph-Based Unsupervised Methods

We propose a new unsupervised method for topic detection that automatically identifies the different facets of an event. We use pointwise Kullback-Leibler divergence along with the Jaccard coefficient to build a topic graph which represents the community structure of the different facets. The problem is formulated as a weighted set cover problem with dynamically varying weights. The algorithm is domainindependent and generates a representative set of informative and discriminative phrases that cover the entire event. We evaluate this algorithm on a large collection of blog postings about different news events and report promising results.

• Type: Conference
• Year: 2008
• Where: COLING
• Authors: Pradeep Muthukrishnan, Joshua Gerrish, Dragomir R. Radev
• Keywords: COLING 2008, Computational Linguistics, Jaccard Coefficient, Pointwise Kullback-leibler Divergence, Set Cover Problem