Abstract
The entity-relationship model depicted in Fig. B.1 serves the purpose of bringing a unified view to the task of extracting degree information from texts. It facilitates the engineering approach of incorporating the proposed representations and methods into other semantic interpretation modules and knowledge bases for text understanding. Different parts of this model have been argued for in different parts of this book.
Classes and their members, the objects, are available in every knowledge base, and each class may have arbitrarily many members. Each object may have several degrees for different graded properties, e.g., height or velocity. Absolute comparisons hold between a degree of an object and a class norm. Class norms are attached to classes and denote typical heights or velocities etc. for expressions like “tall”, “short”, “fast”, and “slow”. Relative comparisons take place between two degrees. Intercorrelations describe patterns of intercorrelation comparisons which may hold between pairs of class norms. Intercorrelation comparisons, absolute comparisons and relative comparisons exhaustively form the class of general comparison relations. They may all use a modifier distance in order to describe expressions like “much taller” or “very tall”. Furthermore, general comparison relations are typed by categories as has been described in Section 7.1.4. Note here that relations that hold for a type, e.g., Comparison-Relation, are also inherited by its subtypes, e.g., Intercorrelation-Comparison, Absolute-Comparison, and Relative-Comparison.
Distance-compares-relations describe linguistic ordering knowledge about distances like “much”, “very”, or “somewhat”. Deduction rules work on triples of comparison relations in order to propagate consequences.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Rights and permissions
Copyright information
© 1999 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
(1999). The Entity-Relationship Model. In: Grading Knowledge. Lecture Notes in Computer Science(), vol 1744. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-46618-5_9
Download citation
DOI: https://doi.org/10.1007/3-540-46618-5_9
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-66934-0
Online ISBN: 978-3-540-46618-5
eBook Packages: Springer Book Archive