Abstract
One of the great challenges to automating open source intelligence from sources such as social media is being able to resolve named entities to identify threat actors, and predict their future activities. In this context, a named entity refers to a person, place, or thing by its name. For example, my name is "Paul Watters," and I belong to a category called "human," which is a subset of "animal," In more general cases, named entity resolution involves two separate stages: resolving the name, and resolving the category according to an ontology. In this chapter, two simple cases are explored which make named entity resolution a difficult task: resolving lexical items within a semantic context, and how the propagation of errors can impact the accuracy of subsequent lexical analysis.
Original language | English |
---|---|
Title of host publication | Automating Open Source Intelligence |
Subtitle of host publication | algorithms for OSINT |
Editors | Robert Layton, Paul A. Watters |
Place of Publication | Waltham, MA, USA |
Publisher | Elsevier |
Chapter | 2 |
Pages | 21-36 |
Number of pages | 16 |
ISBN (Print) | 9780128029169 |
DOIs | |
Publication status | Published - 2016 |
Externally published | Yes |
Keywords
- Computational models
- Errors
- Lexical processing
- Polysemy
- Semantics