Where pattern recognition fits
Most pattern-recognition applications often operate in conjunction with existing data repositories and content management systems to add structure to unstructured data, such as e-mail messages and field intelligence reports. This work is analogous to Web search engines that provide links to relevant documents housed in servers that may be scattered worldwide.
Back-end data repositories associated with pattern-recognition applications can aggregate information from many sources to make the data appear to end users as if it comes from a single source.
Some pattern-recognition programs create repositories of metadata, which essentially is data about data. By classifying incoming information this way, pattern-recognition software can create indexes of subjects in a first step to finding trends and hidden associations in the information. From there, the systems can feed information to user interfaces, such as Web portals or custom-written analytical programs.
In the future, pattern-recognition applications will rely more heavily on distributed processing architectures and high-speed network connections to accommodate massive amounts of data, including large video files. User interfaces dominated by visual-based navigation systems rather than text-based ones will help security analysts more quickly sort through this growing glut of information.
Alan Joch is a freelance writer based in New Hampshire.