SENTENCE-LEVEL EVENT CLASSIFICATION IN UNSTRUCTURED CONTEXTS

Abstract

In this article, we investigate the use of statistical methods for identifying the sentences in a document that describe one or more instance of a specified event type. We treat this task as a text classification problem where each sentence is either classified as one that contains an instance of the target event or one that does not. Event classification at a sentence level is a very challenging task. We estimate a series of uni- gram new models using three essential smoothing approaches, and investigate their overall behavior on classification performance.