A General-Purpose Approach to Temporal Event Ontology Creation
Abstract
One of the major challenges for modern data scientists is providing structure to data. Textual data is especially difficult to interpret and categorize. Much of the meaning found in this natural language data, like news articles or tweets, is contextual and potentially non standard. Attempts have been made to manually create organizational ontologies, but this is usually limited to specialized sub-domains, as the task of providing a complete structure "by hand" across larger domains is unmanageable. We propose a general purpose approach to event ontology creation, building upon a subevent classifier already developed in the initial stage of our research. In this work, we extract events from textual data and create a graph structure showing temporal relationships, using semantic and syntactic methods. This event ontology facilitates faster and more accurate automated data interpretation by providing a structure to textual data. The next stage in the "big data" phenomenon is not accumulating more data, but fully utilizing the vast amount of data already available. Event ontologies are a necessary step in this direction.
Citation
Badgett, Allison Ruth (2017). A General-Purpose Approach to Temporal Event Ontology Creation. Undergraduate Research Scholars Program. Available electronically from https : / /hdl .handle .net /1969 .1 /177577.