Delivering Accurate Media Analysis – Quickly and Cost-Effectively

Perception Metrics delivers accurate media analysis quickly and cost-effectively. Our approach is based on over 10 years of product development in the text analysis industry.

We use a full-text parsing system, which lets us track multiple pieces of information about every word in a body of text. This means that, like the human mind and unlike almost every other player in our market, we can rely on actual grammatical connections between concepts in text. It allows us to get at the same kind of nuanced ideas a reader could, but with the perfect reliability and speed of a computer. We combine this with dictionaries containing tens of thousands of words and phrases in context, including idioms and common expressions, to find out what is being said about you, your competitors, and your market.

Perception Metrics – Benefits to Our Approach

• Results are immediate and efficient – you won’t have to wait for hand-coding of text
• No sampling necessary – we can analyze ALL of your text
• Grammatical connection means we know what is being said about you, not just around you
• Unmatched consistency and reliability – using an automated system means that the same words and phrases will be coded the same way, every single time

Our Approach – Automated Text Analysis

Profiler Plus, developed by Social Science Automation, Inc. in partnership with Ravenbrook, LTD, is a general purpose information extraction system. This technology and subsequent software was borne out of research and application for various government agencies. After ten years of development and support for government, Profiler Plus is now also being used to support more broad applications in the areas of media analysis, campaign and election media evaluation, athlete achievement, profiling, and forensic psycholinguistics.

Profiler Plus is implemented in Common LISP and, in contrast to most statistically-based Natural Language Processing systems; Profiler Plus is entirely rule-based providing fine grained control and complete process transparency. Profiler Plus performs text coding in a highly flexible, multiple pass process, using a set of “dictionaries”. Results are then interpreted by our expert analysts and/or imported into such data analysis programs as SPSS and SAS.

Profiler Plus basic language models for both English and Arabic are available along with a wide range of coding schemes developed for particular information extraction tasks, primarily for the analysis of psychological constructs and of media presentation of people and issues. Profiler Plus accepts text documents as input and generates output in a variety of formats, including, XML, HTML, TXT, LISP, and MDB (MS Access).