Title | ||
---|---|---|
Cross-Language IR at University of Tsukuba: Automatic Transliteration for Japanese, English, and Korean |
Abstract | ||
---|---|---|
This paper describes our cross-language informa- tion retrieval system for the NTCIR-4 CLIR task. Our system, which follows the query translation approach, uses a compound word translation and translitera- tion. Transliteration is effective if a query includes foreign words, such as technical terms and proper nouns, spelled out by phonetic alphabets. We ap- ply our method, which was originally proposed for Japanese Katakana words, to Korean Hangul words and realize JEK transliteration in a single framework. We produced a transliteration dictionary for Japanese and English letters via the Roman representation. To produce a new dictionary, we use the Unicode system to romanize Korean words. We also show the effective- ness of our method by means of experiments. |
Year | Venue | Field |
---|---|---|
2004 | NTCIR | SKATS,Computer science,Compound,Speech recognition,Artificial intelligence,Natural language processing,Unicode,Proper noun,Hangul,Katakana,Transliteration |
DocType | Citations | PageRank |
Conference | 7 | 0.69 |
References | Authors | |
12 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Atsushi Fujii | 1 | 486 | 59.25 |
Tetsuya Ishikawa | 2 | 226 | 30.46 |