Abstract
When performing environment scanning, organizations typically deal with a numerous of events and topics about their core business, relevant technique standards, competitors, and market, among many others, where each event or topic to monitor or track generally is associated with many news documents. To reduce such information overload and information fatigues when monitoring or tracking events, it is essential to develop an effective event episode discovery mechanism to organize all news documents pertaining to an event of interest. In this thesis, we propose a new metric, referred to as TF□Enhanced-IDFTempo and develop an event episode discovery technique that uses the TF□Enhanced-IDFTempo metric as its feature selection method and document representation schemes. Using the traditional TF□IDF and the TF□IDFTempo as performance benchmarks, our empirical evaluation results suggest that our proposed TF□Enhanced-IDFTempo technique outperforms its benchmarks in cluster recall and cluster precision and demonstrates its better capability in reaching the true number of episodes.