As America fights terrorism through its homeland security efforts, the capacity of its intelligence agencies to capture and generate information has far exceeded their ability to analyze it. Intelligence information in the form of text documents floods intelligence and other government agencies. These electronic documents exist in many forms, including cables, field analyst reports, all-caps message traffic, ship manifests, emails, open source documents and other data feeds. Until now, automated text analysis systems have required significant human intervention and have been limited to coding, indexing, keyword search or simple entity extractions that are inadequate because of imprecise and unpredictable results. Furthermore, these systems have been unable to fully leverage the breadth of analytic tools designed for dealing with structured information.
Attensity Intelligence Analysis provides the capability to automate the extraction of "relational events" from free-form text - not only who or what, but why, when, where and how. The solution enables analytical processing by automating the transformation of written language into structured, relational data. The result is dramatically faster and more comprehensive detection of trends, anomalies, patterns and linkages, and more time for analysts to respond to them. Once extracted in this structured form, the information can be pushed downstream to feed virtually any system that processes relational tables. Such tabular data can be stored in a shared data warehouse repository for data mining or link analysis purposes; it can be input directly to a link analysis tool such as Visual Analytics' Visual Link, i2's Analyst's Notebook or other analytical applications, or it could be sent via an automated email alert to an analyst.
Attensity Intelligence Analysis-
- Replaces the slow, expensive, error-prone task of manual document analysis and tagging
- Generates tabular, analysis-ready data structures from text sources up to 20,000 times faster than a human analyst
- Provides robust handling of "noisy" input, such as poor grammar, misspelled words, shorthand and unknown terms
- Integrates easily with existing data sources and targets
The DOD Logistics sector has a long tradition of product and process quality consciousness. Extensive quality processes and methodologies are adopted within internal product planning and manufacturing operations. However, it remains difficult to accurately assess potential quality and service issues after vehicles, equipment and parts and other items critical to our military have been shipped.
The reason for this shortcoming is that the use of standardized labor and part codes to identify reported problems does not correlate well with complex vehicle and component problems, and therefore often result in problem misrepresentation. Fully descriptive text is the most effective medium for communicating problems relating to complex parts and failure chains. Although free-form text entered by technicians and other service personnel is frequently available, it is difficult to analyze in an automated manner.
Attensity offers Attensity Defense Logistics Analysis, a scalable text extraction solution that combines high levels of precision and recall to enable DOD agencies and armed forces to prevent recurring field problems, and enable continuous process improvement. Government logistics organizations utilize Attensity software to improve the reliability and supportability of a variety of military vehicles, weapons and components, by converting unstructured data from service notes and repair logs into relational tables, so that an automated approach can be used to quickly identify and resolve manufacturing and maintenance issues.
Capabilities and benefits of Attensity Defense Logistics Analysis include:
- Extraction of detail from service records which provides unparalleled insight into root cause analysis of problems
- Extraction of information from surveys and registration documents to gauge satisfaction levels and provide alerts to potentially serious design or assembly issues
- Detection of complex cause-and-effect trends involving vehicle systems, sub-systems and failure modes
- Creation and submission of product failure and other adverse event reports
- Event monitoring and notification through integration with existing workflow and alert systems
|
|
