Detection of Terrorist Preparations by an Artificial Intelligence Expert System Employing Fuzzy Signal Detection Theory
Abstract
Terrorist actions are often preceded by preparatory activities that are at least slightly anomalous compared to the activities of the general population (such as unusual purchasing patterns). Analysis of patterns of such anomalies will detect potential preparations for terrorist activity. However, each such anomaly is what is known technically as a fuzzy signal : each anomaly taken individually might indicate terrorist preparations, but usually does not. This paper describes the hypothetical computer system FUSEDOT (FUzzy Signal Expert system for the Detection Of Terrorism preparations). As the name suggests, FUSEDOT applies artificial intelligence expert system technology to the fuzzy signals presented by certain anomalous data, such as interpersonal relationships, financial relationships, travel patterns, purchasing patterns, patterns of Internet usage, and personal background. The components of the system data acquisition, data processing, report generation, and feedback and further system development are described in outline. System limitations are described. FUSEDOT is compared with other systems for the analysis of massive amounts of data to detect terrorism, specifically the Novel Intelligence from Massive Data program, and the Terrorist Information Awareness program. In comparison with these programs, FUSEDOT will implement some superior technology, will be easier to develop, and poses fewer privacy concerns. NATO member nations would do well to consider the development of FUSEDOT, which can be a valuable asset in the defence against terrorism.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 25, 2004
- Accession Number
- ADA460204
Entities
People
- Mark E. Koltko-rivera