Abstract: "We consider the problem of learning a context-free grammar from examples. In this paper, the problem is slightly different from the usual grammatical inference problem. The problem is to learn a context-free grammar adequate for bottom-up parsing or designing [a] bottom-up parser. Ou...

Ausführliche Beschreibung

Bibliographische Detailangaben

1. Verfasser:	Sakakibara, Yasubumi (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	Tokyo, Japan 1988
Schriftenreihe:	Shin-Sedai-Konpyūta-Gijutsu-Kaihatsu-Kikō <Tōkyō>: ICOT technical memorandum 506
Schlagworte:	Parsing (Computer grammar)
Zusammenfassung:	Abstract: "We consider the problem of learning a context-free grammar from examples. In this paper, the problem is slightly different from the usual grammatical inference problem. The problem is to learn a context-free grammar adequate for bottom-up parsing or designing [a] bottom-up parser. Our final goal is to present the system using grammatical inference methods to develop a grammar for bottom-up parsing. Then what do we mean by 'a grammar adequate for bottom-up parsing'[?] Our answer is that the grammar should have the intended structure for parsing and allow the process of bottom-up parsing to be made easily Furthermore for a practical use, we require that the grammar should be learned from positive-only examples and the grammar should be learned efficiently. To achieve those requirements, we present an efficient algorithm for learning a context-free grammar from positive examples of structural descriptions. Structural descriptions of a context-free grammar are unlabelled parse trees of the grammar, the shapes of parse trees. Thus the input to the learning algorithm is a finite set of shapes of parse trees. We show that the learning algorithm learns a grammar which is structurally equivalent to the unknown grammar and achieves the polynomial time bound.
Beschreibung:	17 S.