Horizon-Scanning Experiments on Scientific Literature

• Activity Number: 608
• Type: Topic Contributed
• Date/Time: Thursday, August 5, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Defense and National Security
• Author(s): Avory Bryant*+
• Companies: Navy
• Keywords: Horizon Scanning ; Information Visualization ; Information Science ; Data Mining

Abstract

Increased access to vast amounts of scientific research online has spured great interest in various areas of Information Science in exploiting these new reserviors of data. "Scientific Literature" is especially of interest due to the attention to detail afforded to metadata, who/when/where/what, by researchers or those who redistribute their work. In this work we attempt to develope several Horizon Scanning signatures that could be used to identify "interesting" trends and/or abnormalities within the data. "Interesting" in regards to whom? In our case we worked closely with scientific literature analysis to develope potential signatures. Signatures are calculated using a variety of statitstical data analysts techniques such as time series analysis, odds ratios, clustering, and outlier detection. We also provide methods for visualizing the data with signature results.