A New Method of Evolution Event Tracking Based on Variable Query

As a feature co-occurrence approach, the traditional event tracking technique encounters a dilemma when setting the tracking similarity threshold that can diminish its tracking effectiveness. In this paper we present a new event tracking method-variable query, which can automatically adjust query terms and its corresponding weights according to the event evolution stages and tracking similarity, and solve the dilemma well. Experiment shows that the method of variable query can improve the precision and decrease the false rate of tracking.