Novel Topic Authorship Attribution
Abstract
The practice of using statistical models in predicting authorship (so-called author-attribution models) is long established. Several recent authorship attribution studies have indicated that topic-specific cues impact author-attribution machine learning models. The arrival of new topics should be anticipated rather than ignored in an author attribution evaluation methodology; a model that relies heavily on topic cues will be problematic in deployment settings where novel topics are common. In order to effectively deal with novel topics, we create author and topic vectors and attempt to project out the topic influences from each document. Although our experiments did not validate our assumptions, they do point out a possible problem with a common assumption in authorship attribution research.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2011
- Accession Number
- ADA543929
Entities
People
- Randale J. Honaker
Organizations
- Naval Postgraduate School