Automatic Detection of Temporal Relations in Text
Name
Taaniel Saarnik
Abstract
The aim of this thesis is to create machine learning models that identify events described in newspaper texts and identify temporal relations between events and temporal expressions. Newspaper texts contain many different events that can be interpreted as being connected via temporal relations. To be able to use that information in different natural language processing tasks, it is crucial to tag these words somehow. Manual tagging is very troublesome and time-consuming task and that is why automatic system would be very beneficial. This thesis describes earlier solutions to this problem and a process of creating machine learning models to solve this.
Graduation Thesis language
Estonian
Graduation Thesis type
Bachelor - Computer Science
Supervisor(s)
Siim Orasmaa
Defence year
2021