Name
Abstract | ||
|---|---|---|
This paper describes our retrieval system for the NTCIR-3 CLIR task, focusing on Japanese and En- glish. We integrate query and document translation methods to improve retrieval accuracy, and perform clustering to improve browsing efficiency. In query translation, to derive possible translations for user queries, we use dictionaries and perform a transliter- ation method, which generates translations for out-of- dictionary loanwords based on the Japanese phono- gram system. We also use a probabilistic model to re- solve translation/transliteration ambiguity to improve the query translation accuracy. We show the effective- ness of our system with respect to the NTCIR-3 J-J, E-E, J-E, and J-JE subtasks. |
Year | Venue | Keywords |
|---|---|---|
2002 | NTCIR | ntcir-3,cross-language information retrieval,transliteration,query translation,probabilistic model |
Field | DocType | Citations |
Query language,Information retrieval,Query expansion,Computer science,Artificial intelligence,Natural language processing,Statistical model,Cluster analysis,Ambiguity,Phonogram,Transliteration | Conference | 1 |
PageRank | References | Authors |
0.38 | 12 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Atsushi Fujii | 1 | 486 | 59.25 |
| Tetsuya Ishikawa | 2 | 226 | 30.46 |