E.D. Liddy (USA)
Information Extraction, Natural Language Processing, Subjectivity, Affect, Certainty
This paper reports on techniques for extending Information Extraction capabilities beyond the recognition, tagging, and extraction of entities, events, and the relations amongst them that are reported in text to capture and represent the subtler aspects of content, whether in narratives, factual reports, or opinion pieces. As statements in text often exhibit subjective colorations that can be detected, analyzed, and interpreted by NLP algorithms for presentation to users for a more accurate understanding of what might otherwise be represented as straightforward information. This paper focuses on the temporal, certainty, and affective aspects, each of which have shown promise for greater sophistication in terms of what Information Extraction systems can glean from text.
Important Links:
Go Back
