Finite-state parsing and disambiguation

    Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

    Abstract

    A language-independent method of finite-state surface syntactic parsing and word-disambiguation is discussed. Input sentences are represented as finite-state networks already containing all possible roles and interpretations of its units. Also syntactic constraint rules are represented as finite-state machines where each constraint excludes certain types of ungrammatical readings. The whole grammar is an intersection of its constraint rules and excludes all ungrammatical possibilities leaving the correct interpretation(s) of the sentence. The method is being tested for Finnish, Swedish and English.
    Original languageEnglish
    Title of host publicationProceedings of the 13th conference on Computational linguistics, Vol. 2
    EditorsHans Karlgren
    Number of pages4
    Place of PublicationHelsinki
    Publication date1990
    Pages229-232
    ISBN (Print)952-90-2026-0, 952-90-2028-7
    Publication statusPublished - 1990
    MoE publication typeA4 Article in conference proceedings
    EventThe 13th International Conference on Computational Linguistics - Helsinki, Unknown
    Duration: 20 Aug 199025 Aug 1990
    Conference number: 13

    Fields of Science

    • 612 Languages and Literature
    • natural language
    • syntactic parsing
    • finite-state

    Cite this