Natural language IAS: style metrics from semantic analysis
Lauren M. Stuart, Julia M. Taylor, Victor Raskin · Annual Information Security Symposium · 2015
Stylometry is the quantification of author style such that authorship of a text can be posited, verified, or obfuscated. Style features currently in use capture the surface features of texts (such as punctuation use, misspellings, words or parts of words, and morphology), but some qualities of author style may be better captured by, or in conjunction with, meaning-based features. This poster outlines ongoing work in positing and evaluating author style quantification using meaning representation structures.