This is experimental code for a similarity-based personal identification program, which has many uses in security, healthcare, and usability features. The method relies on cosine-based word similarity matrices, generated from personnel in the Enron dataset. Preliminary results are good. The methods are kludgy, its really just the similarity metrics that are being demo'ed.
-TODO: add Enron resource data -improve metrics -more experiments to determine useful projections of similarity data
All work in this repo is copyright Jesse Waite, 2014, 2015, 2016