Using Natural Language Processing for Identifying and Interpreting Tables in Plain Text
Shona Douglas, Matthew F. Hurst, David I. Quinn · 2007
Characteristics of Tables 2.1 Table Denotations A simple table in some abstract sense denotes a relation among sets of values. 2 (We can think of this as corresponding to relations or views in relational database terms. See, for example, [Ull88]) In these terms, the underlying representation of a table is a set of n-tuples, where n is the number of domains or value sets in the table. The n-tuples of the table is that subset of the cross-product D 1 \\Theta D 2 \\Theta : : : \\Theta Dn for which the relation holds. This is the canonical form of the table. For our nlp application, we need a relatively detailed ontology or world model of the objects of the sublanguage of construction. The classes of this world model we call world domains; they include 1 We thank Stephen McCarron of BICC. We acknowledge the support of the Department of Trade and Industry, the Engineering and Physical Sciences Research Council, and the BICC Group on the CISAU project (IED4/1/5818), and the Human Communi...