2006 Fiscal Year Final Research Report Summary
Multi-document Summarization by Taking into account the relation between documents
| Project/Area Number |
16300041
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
OKUMURA Manabu Tokyo Institute of Technology, Precision and Intelligence Laboratory, Associate Professor, 精密工学研究所, 助教授 (60214079)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
NANBA Hidetsugu Hiroshima City University, Faculty of Information Science, Associate Professor, Associate Professor, 情報科学部, 講師 (50345378)
|
|---|
| Project Period (FY) |
2004 – 2006
|
| Keywords | Multi-document summarization / Genre identification / cross-document sentence relationships / opinion sentence extraction |
|---|
| Research Abstract |
In this research, we develop a multi-document summarization technique than can cope with a document collection of multiple genres. Generally, in multi-document summarization system, it is necessary to collect a text collection to be summarized as a preprocessing. In that process, documents of multiple genres might be grouped as a collection. We need to develop different summarization techniques for collections of different genres. Therefore, in this work, targeting at the following three genres. Web pages, scientific papers, and newspaper articles, that can be considered to be the major genres now, we realize a multi-document summarization technique which can cope with a document collection of multiple genres, by developing the following sub-modules : (1) a module of genre identification, (2) a multi-document summarization module for a document collection of genre, and (3) a module of integrating the summaries produced by the module (2) and generating the final summary.
|
