Specifications on digital hardware systems typically contain descriptions and requirements expressed in natural language and diagrams of various types. The objective of the research reported here is the automatic detection of common references ("coreferences") to objects in natural language specification statements in order to permit automatic integration of requirements. This thesis describes a prototype system for detecting coreferences. First, the natural language statements are translated into conceptual graphs (semantic nets). Then, these graphs are scanned by a rule-based system to determine whether each concept that is encountered is the definition of a new concept or a reference to a previously defined concept. Tests performed on the system developed indicate a high percentage rate of correct classifications. / Master of Science
Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/42360 |
Date | 29 April 2009 |
Creators | Shankaranarayanan, S. |
Contributors | Electrical Engineering, Cyre, Walling R., Armstrong, James R., Gray, Festus Gail |
Publisher | Virginia Tech |
Source Sets | Virginia Tech Theses and Dissertation |
Language | English |
Detected Language | English |
Type | Thesis, Text |
Format | viii, 116 leaves, application/pdf, application/pdf |
Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Relation | OCLC# 30505852, LD5655.V855_1994.S526.pdf |
Page generated in 0.0016 seconds