发明名称 |
SIMILARITY DISCRIMINATING METHOD |
摘要 |
PROBLEM TO BE SOLVED: To provide a similarity discriminating method capable of discriminating similarity by calculating the similarity by combining plural word bases even when two words the similarity of which is discriminated are not included in one word base in the plural word bases. SOLUTION: Words commonly registered in the word bases 1, 2 are retrieved as a word set α (step S15), the similarity between individual words in the word set α and a word (a) is calculated by using a word base 1 (step S17), words with high calculated similarity are retrieved from the word set α and defined as a word set β (step S19), the similarity between individual words in the word set β and a word (b) is calculated by using the word base 2 (step 21) and the similarity between the words (a) and (b) is calculated by using the similarity of the word (a) and the similarity of the word (a) which are calculated as above (step S23). |
申请公布号 |
JP2001243230(A) |
申请公布日期 |
2001.09.07 |
申请号 |
JP20000049948 |
申请日期 |
2000.02.25 |
申请人 |
NIPPON TELEGR & TELEPH CORP <NTT> |
发明人 |
INAGO KIBO;KASAHARA KANAME |
分类号 |
G06F17/21;G06F17/22;G06F17/27;G06F17/30 |
主分类号 |
G06F17/21 |
代理机构 |
|
代理人 |
|
主权项 |
|
地址 |
|