IMC's Advanced Meaning-Based Search
IMC commonly deploys this combination of capabilities for law enforcement and intelligence agencies. Semantic analysis begins with finding the right data and documents to analyze. IMC's Advanced Search is intelligent software that can index digital data archives in multiple languages and multiple forms – text, image, audio, video — so you can search for relevant documents in them. Using natural language queries, you can then search those archives as well as ones indexed using other methods.
Automatic Summarization can then create individual summaries of the content in direct hit documents that Search presents in a relevance-ranked list. With each successive search session, Advanced Search also "learns" how to perform more relevant searches based on your previous searches. As you accept and reject search hits, the software increasingly refines the scope of the search and does not retrieve documents it considers irrelevant. It also offers predefined libraries that reduce the effort involved in accomplishing multi-session complex searches and automatically suggest alternative search strategies.
Less robust and complex search tools provide Google-like lists of hits that provide snippets of text from documents that users click on to access the entire document and read through it to determine its relevance, which slows searching and wastes valuable investigative time.
We also often deploy Advanced Search in concert with Automatic Linking to link documents you have located to others that this capability "thinks" are related. Automatic Lnking focuses on related documents that you might not have imagined were useful for your present purpose.
We can also add categorization and "push" mechanisms so software pushes documents it determines might have possible relevance and automatically categorizes them into user-defined categories displayed in windows on the user's interface. Push also automatically alerts you to new documents in a category.
These capabilities gain additional synergy when deployed in an enterprise portal environment that integrates disparate repositories and offers an administrator- or user-configurable interface with flexible categories. Participants can share their data in common knowledge bases containing investigators' findings on a case or ongoing investigations.
For instance, if an intelligence officer were gathering data about a suspected terrorist, he might define various category windows on his portal interface. One might display all direct hits from external websites on documents where the suspect's name appears; another the results of auto-linking from the data in the suspect's phone records -- like the names of people called in the last year — to information about those people in his own case management system. Another might display the results of auto-linking from data in the suspect's credit card transaction records like names of chemical retailers listed there to documents in related agencies' repositories where the names of those companies appear.
Connections among the various data sources may become evident to the officer so that he can extrapolate about other terrorists in the suspects cell as well as possible businesses funding and laundering funds in support of his activities.
It should be mentioned that most of our tools are also certified for use in various intelligence and law enforcement agencies that deal with classified data.