|March 18, 2010 |
Human Factors in Text Analytics: InStream Media
InStream Media, one of Stevens spinout companies, has developed a sophisticated suite of algorithms and Internet applications based on psycho-linguistics, information theory, and advanced text mining that can identify deceptive text contained in emails, social networking sites, blogs, SMS, advertisements, etc. Accuracy rates are typically over 96%.
This is a revolutionary new approach to text mining and is available for demonstration at www.stevens.edu/deception. The InStream Team is in the process of building a broad set of Internet based applications including widgets, plug-ins and apps for social networking sites, web browsers, email programs, word processors and mobile devices.
The company sees this product as a key component in helping businesses and individuals protect themselves against the ever-growing numbers of scam artists, cyber stalkers, misrepresented advertisements and a host of other communications that are not being truthful and/or significantly stretching reality. The market for this product is enormous and InStream Media’s Cyber Security Division is a leading player in accuracy rates with low false positives.
The Company plans to launch an iPhone application in the second quarter of 2010 and is also in conversations with potential OEM suppliers to add this service to their core products such as e-mail programs.
Check back soon for an in-depth discussion of InStream and the technology that makes it possible in an upcoming Research Article!
For more information, please contact Malcolm Kahn.