SBIR Phase II: Software to Aggregate, Correlate, Analyze and Trend data for Knowledge Management in Decision Making

This Small Business Innovation Research (SBIR) Phase II project addresses the challenges for entities seeking to derive reliable and actionable information from enormous quantities of online ""chatter"" (online content from a variety of sources such as blogs, industry-focused sites, and media-generated material). Phase II will focus on technical objectives that will enhance the quality and reliability of the information produced by the ChatterSpike concept researched in Phase I. These objectives fall into three categories: data cleansing, context analysis, and basic commercial readiness. Their achievement will require the design, development and implementation of novel, niche-focused algorithms that will enable the mining and evaluation of thousands of online sources and the production of data with quantifiable quality metrics relating to authority, reliability, influence, and sentiment. The resulting product will algorithmically determine and quantitatively measure and evaluate these parameters in real time as it mines online sources for data, validating its conclusions and re-validating them every time it performs a retrieval operation. By focusing on specific industry niches, the technology produced will enable the production of automated, highly tailored, detailed reports with a high degree of quantitatively-confirmed reliability. This capability will result from the creation of novel algorithms designed to exploit cutting-edge theoretical approaches to extracting, validating, and evaluating information from a multiplicity of online sources. These reports will be superior to the manual reports produced by currently available technologies and approaches. In addition, if successful, the technology will have significant societal benefit. Companies will be able to react more quickly to meet consumer demands and to correct negative trends in consumer opinions. The technology will also be able to detect trends reliably at a very early stage; in some cases weeks or months before they become obvious and are detected by other methods.