| dc.contributor | 英文系 | en |
| dc.creator (作者) | Chui, Ka-Wai | en |
| dc.creator (作者) | Lin, Yia-Ping | en |
| dc.creator (作者) | Huang, Shuanfan | en |
| dc.date (日期) | 1989-10 | - |
| dc.date.accessioned | 31-Oct-2018 14:00:25 (UTC+8) | - |
| dc.date.available | 31-Oct-2018 14:00:25 (UTC+8) | - |
| dc.date.issued (上傳時間) | 31-Oct-2018 14:00:25 (UTC+8) | - |
| dc.identifier.uri (URI) | http://nccur.lib.nccu.edu.tw/handle/140.119/120817 | - |
| dc.description.abstract (摘要) | National Taiwan University This paper aims to propose a new parsing strategy to tackle the notorious prepositional-phrase attachment problem (PP attachment problem) in our NTUMT system. First of all, correct PP attachment is determined in our PP-Attachment Table (PAT), which requires both syntactic and semantic analyses of verbs, nouns and prepositions in lexicon. PAT is indeed the component where all the idiosyncratic attachment conditions are specified for prepositions. As to our parsing strategy, it can be considered as an interaction between two drivers-Intelligent ATN (IATN) and Phrase Structure ATN (PATN). IATN scans the input sentence leftwards and activates PATN to construct the first bar-level structures. PATN is then responsible for building up structures whenever IATN gives the command. Correct PP attachment is governed by the seven states in the IATN Grammar, giving priority to verbs. Since our system only outputs one parsing tree, for a PP which is ambiguous in attaching both to the preceding verb and the preceding noun, the PP will be assigned to be a verb modifier. | en |
| dc.format.extent | 122 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.language | en_US | en |
| dc.relation (關聯) | Journal of Inforamtion Science and Engineering, Vol.5 No.4, pp.449-460 | en |
| dc.subject (關鍵詞) | PP-attachment aable; attachment conditions; parsing strategies; intelligent ATN; phrase structure ATN | en |
| dc.title (題名) | NTUMT strategy for prepositional-phrase attachment | en |
| dc.type (資料類型) | article | en |
