Bilkent University
Department of Computer Engineering


Discovering Story Chains: A Framework Based on Zigzagged Search and News Actors


Çağrı Toraman
PhD Student
Computer Engineering Department
Bilkent University

A story chain is a set of related news articles that reveal how different events are connected. This study presents a framework addressing several research questions to discover story chains, given an input document, in a text collection. The framework has three complementary parts that a) scans the collection, b) measures the similarity of chain-member candidates with the chain, and c) measures similarity among news articles. In the scanning part, we apply a novel text-mining method that uses a zigzagged search that reinvestigates past documents based on the updated chain. We also utilize news actors in a social network to reflect direct and hidden connections among events. The performance of the framework is evaluated by two user studies in terms of four effectiveness measures—relevance, coverage, coherence, and ability to disclose relations. The first user study compares several versions of the framework by varying parameters to find a highly-effective version. The second one compares the highly-effective version with three baseline algorithms. The results show that the effectiveness of the framework algorithm is not statistically worse than any baseline, while we statistically improve the effectiveness in 67% cases.


DATE: 28 November, 2016, Monday @ 16:00