Using Fuzzy Logic in Evaluating User Tabled Correlation Rules for COMINT
Abstract
This paper presents a methodology for performing level-1 fusion involving organizational structures and for determining which structures, if any, can be recognized. Available data consists of contact reports received via various communications circuits. Information for correlation arises from three main sources: parametric signature data such as radio frequency, nominative categorical information such as unit identifier, and location information such as intercept position. Both individual and aggregations of such data separately give rise to correlation sub-scores. Each sub-score provides evidence about whether a new report correlates with an existing entity. The paper identifies methods and reports some preliminary results correlation, which use the operators of fuzzy logic, and adaptive weighting to enforce analysts specified correlation rules.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 17, 2000
- Accession Number
- ADA392900
Entities
People
- John M. Palmer