A Framework for Temporal Resolution

Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC 2004)

Abstract

The ability to identify and analyse temporal information is important for a variety of natural language applications, such as information extraction, question answering and multi-document summarisation. This paper discusses the automatic annotation of temporal information in newswire texts. The temporal tagger adheres to a framework initially designed for anaphora resolution and recently adapted for temporal resolution. It achieves an accuracy of 83.66% in identifying and labelling temporal expressions.