| Presentation | 2008-02-07 Extracting Opinions for Broadcasting using Multiple Sequence Alignment Takeshi S. KOBAYAKAWA, Naoto KATO, Jun'ichi TSUJII, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The classification by the computer analysis of electronic opinions for braodcasting is studied. When analyzing the opinions for broadcasting, the characterestic point is that it is hard to enumerate with limited words beforehand what the opinion is about. Therefore, it is necessary to determine the part opinion is expressed without determining the part the opinion is about. In order to determine the part opinion is expressed, not the local n-gram information but a global information is dealt in this study. The frequently used expressions to express opinions are extracted by aligning many opinions which are morphologically analyzed. Taking the alignment more than two sequence is called multiple sequence alignment, which has an application for frequently used expressions. Multiple sequence alignment is only tractable when an approximated method is used, while we propose an approximated method to preserve an ambiguity which is likely to take place in natural language processing. The proposed method essentially becomes an exponential order of computation, so several cutoffs are introduced to guarantee the possibility of the computation. Since there are plural final candidates, external criteria are introduced to order the candidates. The proposed method is applied to the corpus of opinions for broadcasting which were collected in the experiment, and the accuracy of extracting expressions are compared with the conventional method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Multiple Sequence Alignment / Dynamic Programming / Document Classification / Ambiguity of Alignment |
| Paper # | NLC2007-91 |
| Date of Issue |
| Conference Information | |
| Committee | NLC |
|---|---|
| Conference Date | 2008/1/31(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Natural Language Understanding and Models of Communication (NLC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Extracting Opinions for Broadcasting using Multiple Sequence Alignment |
| Sub Title (in English) | |
| Keyword(1) | Multiple Sequence Alignment |
| Keyword(2) | Dynamic Programming |
| Keyword(3) | Document Classification |
| Keyword(4) | Ambiguity of Alignment |
| 1st Author's Name | Takeshi S. KOBAYAKAWA |
| 1st Author's Affiliation | Japan Broadcasting Corporation:Graduate School of Information Science and Technology, The University of Tokyo() |
| 2nd Author's Name | Naoto KATO |
| 2nd Author's Affiliation | Japan Broadcasting Corporation |
| 3rd Author's Name | Jun'ichi TSUJII |
| 3rd Author's Affiliation | Graduate School of Information Science and Technology, The University of Tokyo |
| Date | 2008-02-07 |
| Paper # | NLC2007-91 |
| Volume (vol) | vol.107 |
| Number (no) | 480 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |