Search |  Contact |  SRI Home Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap. Do not follow this link, or your host will be blocked from this site. This is a spider trap.A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A ASRI International.  333 Ravenswood Avenue.  Menlo Park, CA 94025-3493. SRI International is a nonprofit corporation.

Publication Details

Database Editing Metrics for Pattern Matching

by Ruspini, E. and Thomere, J. and Wolverton, M.

in Proceedings of the IEEE International Conference on Computational Intelligence for Homeland Security and Personal Safety (CIHSPS-04) pp. 39-45,

Address: S. Giuliano - Venice, Italy
Organization: IEEE Computational Intelligence Society
Aug 2004.

Abstract

Pattern-matching techniques are important tools to treat problems in several fields, including bioinformatics, case-based reasoning, information retrieval, and pattern recognition. These procedures are important in homeland security and crime prevention applications because the underlying problems require the discovery, in large databases, of instances of patterns known to be associated with illegal activities. While pattern matching may be defined in strict terms as the satisfaction of a logical expression, defining the pattern, by a set of assertions contained in the database, the value of relevant procedures is considerably enhanced by permitting the discovery of approximate matches between database and patterns. The notion of approximate matching is based on the consideration of soft predicates, which may be satisfied to a degree, rather than the conventional crisp predicates of classical logic. This paper introduces a family of metrics to measure the degree of qualitative match between a database and a pattern, that is, an elastic constraint on database objects and their relations. These metrics provide a formal foundation for the application of graph-editing metrics—measures of the cost associated with graph transformations—to pattern-matching problems. The degree of matching between database and patterns is determined by means of similarity measures that gauge the resemblance between pairs of objects. In our treatment, these measures have a semantic basis stemming from consideration of knowledge structures, such as ontologies, describing common properties of two objects. Approximate pattern matching is treated as the process of modifying databases into a transformed database that strictly satisfies the constraints expressed by the pattern. Associated with each transformation is a measure of admissibility derived from the similarity between the original and transformed databases. The degree of matching of database to pattern is defined as the admissibility of the transformation with highest admissibility value.

Electronic Copies


Adobe PDF

BibTeX

EndNote
Copyright ©IEEE

Associated Projects

LAW

The Link Analysis Workbench

The goal of this project is to develop the Link Analysis Workbench (LAW), a Web-accessible tool where analysts and machines collaboratively perform link analysis by defining hierarchical and temporal patterns, that include uncertain and qualitative elements, and by defining search strategies for pattern application, through a graphical user interface that supports direct graphical browsing and editing of patterns, search strategies, and summaries and details of resulting matches.

AIC Personnel

Name Title E-mail
Ruspini, Enrique H Alumnus
Thomere, Jerome F Computer Scientist
Wolverton, Michael J Senior Computer Scientist

SRI International
©2014 SRI International 333 Ravenswood Avenue, Menlo Park, CA 94025-3493
SRI International is an independent, nonprofit corporation. Privacy policy