Discovering Temporal Relations with TICTAC
Georgiana Pu · 2007
Temporal information plays an important role in many NLP applications. The identiflcation of temporal relations between temporal entities (events and temporal expressions) is indispensable in obtaining the temporal interpretation of a given text. This paper presents our approach for discovering temporal relations using the temporal annotation system we have developed. This system is called TICTAC (Syntactico-Semantic Temporal Annotation Cluster) and it comprises both knowledge based and statistical techniques. It has achieved the best performance among all systems participating at the TempEval competition organised as part of SemEval-2007, competition that evaluated temporal relation identiflcation capabilities.