| Presentation | 2012-02-03 Proposal of an Extraction Method of Multilingual Parallel-text Candidates Using Web Data Taku FUKUSHIMA, Takashi YOSHINO, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In this study, we have proposed an extraction method of multilingual parallel-text candidates for automatic creating of accurate multilingual parallel texts. Multilingual parallel texts are created and collected for multilingual communication. Translators translate example sentences into parallel texts. However, translators was feeling the burden of translation. It is barrier to collect parallel texts. Therefore, we proposed an extraction method of translation candidates. This method uses existing parallel texts, multilingual dictionary, machine translation and Web search engine. Moreover, the purpose of this method gets rid of the burden of translators. The contributions of this paper are the following results: (1) We proposed an extraction method of multilingual parallel-text candidates creating accurate parallel texts. This method includes "a translation candidates creating method" and "a selection method of accurate parallel-text." (2) We proposed the translation candidates creating method for increasing translation candidates including accurate translations. Moreover, we showed that this method created accurate translation candidates about 1.5 times more than general machine translation. (3) We proposed the selection method of accurate parallel-text for selecting accurate translation in translation candidates. Moreover, we showed that the accuracy of this method increased by about 16 points than the accuracy of general machine translation. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | parallel text / machine translation / web search / to evaluate accuracy |
| Paper # | NLC2011-66 |
| Date of Issue |
| Conference Information | |
| Committee | NLC |
|---|---|
| Conference Date | 2012/1/26(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) | Proposal of an Extraction Method of Multilingual Parallel-text Candidates Using Web Data |
| Sub Title (in English) | |
| Keyword(1) | parallel text |
| Keyword(2) | machine translation |
| Keyword(3) | web search |
| Keyword(4) | to evaluate accuracy |
| 1st Author's Name | Taku FUKUSHIMA |
| 1st Author's Affiliation | Graduate School of Systems Engineering, Wakayama University() |
| 2nd Author's Name | Takashi YOSHINO |
| 2nd Author's Affiliation | Faculty of Systems Engineering, Wakayama University |
| Date | 2012-02-03 |
| Paper # | NLC2011-66 |
| Volume (vol) | vol.111 |
| Number (no) | 427 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |