STORING SYSTEM FOR MORPHEME ANALYSIS RESULT INFORMATION ON DOCUMENT READER AFTER-PROCESSOR

PURPOSE:To facilitate the work in the subject system by applying the analysis of morphemes to a candidate character assembly string outputted from a character recognizing device as an after-process, storing the obtained after-process document in an external storage and using the analysis result info...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: TAKAO NORIYASU
Format: Patent
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:PURPOSE:To facilitate the work in the subject system by applying the analysis of morphemes to a candidate character assembly string outputted from a character recognizing device as an after-process, storing the obtained after-process document in an external storage and using the analysis result information including the positions and the grammatical attributes of morphemes as the control information to the after-process result document. CONSTITUTION:A candidate character input part 11 fetches the candidate character assembly string and the evaluation value outputted from a character recognizing device 6 and stores them in a work area serving as a candidate character table of a document reader after-process main body 12. An after-process part 13 performs the after-processes of words which could not be decided by the main body 12, the output of the after-process result document, the storage of the information on the morpheme analysis result, etc. A word collation part 14 collates the words obtained by combining the candidate characters of the candidate character table with a word dictionary 16. A grammar collation part 15 decides the presence or absence of an adjacency enable state between words based on the adjacency information obtained from the dictionary 16. If the adjacency enable state is confirmed, a grammar dictionary 17 is referred to. The information thus obtained can be applied to a natural language processing system.