AAAI Publications, Twenty-Third International FLAIRS Conference

Font Size: 
GPAT: A Genre Purity Assessment Tool
Philip Michael McCarthy

Last modified: 2010-05-06


This study introduces a Genre Purity Assessment Tool (GPAT). GPAT calculates genre purity by using SIF n-graphs (statistically improbable graph strings) to identify genre characteristics in text. The study describes the tool and assesses it across five experiments that feature a variety of text types and text lengths. The results demonstrate that GPAT is at least as effective as a system that uses a combination of 30 complex textual analysis indices. The results further demonstrate that GPAT is informative on texts as short as three words. The study is of value to discourse psychologists, psycholinguistics, and any researchers for whom the genre of texts is a component of the analysis.

Full Text: PDF