1987 Fiscal Year Final Research Report Summary
Comprehensive Machine Dictionary for Natural Language Processing
| Project/Area Number |
60302090
|
|---|
| Research Category |
Grant-in-Aid for Co-operative Research (A)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Informatics
|
|---|
| Research Institution | Kyushu Institute of Technology (1986-1987)
Kyushu University (1985) |
|---|
Principal Investigator |
YOSHIDA SHO Faculty of Comp. and Sys., Kyushu Ins. of Tech.,Professor, 情報工学部, 教授 (80039065)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TANAKA YASUHITO Himeji College , Associate Professor, 助教授 (00163585)
OKADA NAOYUKI Faculty of Engineering, Oita Univ., Professor, 工学部, 教授 (80037837)
TSURUMARU HIROAKI Faculty of Engineering, Nagasaki Univ.,Associate Professor, 工学部, 助教授 (90039678)
NAGAO MAKOTO Faculty of Engineering, Kyoto Univ., Professor, 工学部, 教授 (30025960)
HITAKA TORU Faculty of Engineering, Kyushu univ.,Associate Professor, 工学部, 助教授 (30037931)
|
|---|
| Project Period (FY) |
1985 – 1987
|
| Keywords | Machine Dictionary / Semantic Dictionary / Thesaurus / Natural Language Processing / 自然言語理解 |
|---|
| Research Abstract |
Very few machine dictionaries are freely available in common for natural language processing researchers in Japan. And grammartical and semantic informations of words stored in them are very poor for advanced processing of natural language. So, it is urgent to develop a fully integrated and highly qualitative machine dictionary available for every researcher.
The aim of our project is to discuss,from various point of view, what is adequate semantic information for machine dictionary and to find out fundamentals for developing fully integrated machine dictionary every researcher can use by editting according to his/her purpose.
Achievement of our project is summarized as follows. (1) We examined the form of definition sentences of entry words in 10 general word dictionaries and idiomatic phrases which played important role to explain semantics of entry words were extracted. (2) Semi-automatic procedure was developed to extract important semantic information of entry words, such as superordinate/subordinate and part/whole relations, from definition sentences in general word dictionaries. (3) Desirable structure and role of machine dictionary for natural language understanding systems were clarified and it was shown that several problems so far were solved by using the new machine dictionary. (4) To cope with new words not stored in a machine dictionary, word model which simulates the construction process of words was proposed and the effectiveness of the model was examined. (5) A text database management system available for collecting useful information for a machine dictionary from a large volume text file was developed.
|
Research Products
(11 results)
