Author and Group Insight through Linguistic Expression (AGILE)

A document can express a wealth of information about its author, including potential groups they might sympathize with to the opinions or emotions they convey. This effort advances authorship analysis into deeper inference and understanding about who the author is. Demographic attributes such as gender are predicted, and the range of emotions expressed in their writing are automatically identified and incorporated into their profile. Our capabilities will detect vulnerabilities to the authorship analysis system, including attempts to deceive authorship identification. The AGILE system will produce a multi-faceted portrait of the individual. A visual analysis tool will display results and will integrate information about groups of similar authors. Our authorship analysis will be extended into multiple languages, beginning with Arabic. Despite being one of the world"s most widely spoken languages, sentiment and semantic analysis of Arabic text is a largely underexplored area of research. This software will evolve capabilities for inference on Arabic text. DAC"s authorship analysis efforts will lead to a full-scale portrait of an author, including multi-lingual text analysis and visual tools to help users quickly navigate through a versatile and complex analysis of the author"s identity.