1998 Fiscal Year Final Research Report Summary
Information Extraction and Retrieval from Lange Text Data
| Project/Area Number |
08458081
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Kyushu Institute of Technology |
|---|
Principal Investigator |
NOMURA Hirosato Kyushu Institute of Technology, Department of Artificial Intelligence, Professor, 情報工学部, 教授 (30208392)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
NAGAI Hidetoshi Kyushu Institute of Technology, Department of Artificial Intelligence, Assistant, 情報工学部, 助手 (60237485)
NAKAMURA Teigo Kyushu Institute of Technology, Department of Artificial Intelligence, Lecturer, 情報工学部, 講師 (40198221)
|
|---|
| Project Period (FY) |
1996 – 1998
|
| Keywords | Intelligent Information Access / Information Extraction / Information Retrieval / Information Summarization |
|---|
| Research Abstract |
This research concerns Information Extraction, Information Retrieval, and Information Summarization form a large text data set. The approach is based on a pattern-match processing which utilizes surface characteristics in linguistic representations. This method does not require any heavy linguistic processing and any deep analysis of semantic information while it results in high quality and high speed information processing for Information Extraction, Information Retrieval, and Information Sunmarization.
First of all, we elaborated on a Dialogue System for Information Retrieval. It is possible that a user's request is vague and unclear. We investigated a method for making clear the user's request by providing a cooperative and friendly navigation agent and by incorporating a processing strategy which applies a fuzzy calculation for disambiguation.
Second of all, we studied Information Extraction from News Articles. We investigated several useful strategies for designing templates for pattern-matching. We actually developed a large set of templates from 2000 news articles concerning new products.
Last of all, we analyzed linguistic characteristics of sentence endings and then proposed a semantic model for sentence types. By applying this investigation, we proposed a strategy for producing a text summarization by eliminating unimportant sentences and then combining remaining sentences as an article.
All of the experimental systems developed by ourselves are ready for demonstrations on the Web at out HomePage on the internet.
|
