µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
¿ÂÅç·ÎÁö ±â¹Ý ÀǹÌÀû À¯»ç¼º °Ë»ö ±â¹ý |
¿µ¹®Á¦¸ñ(English Title) |
Semantic Similarity Search based on Ontology |
ÀúÀÚ(Author) |
±è±â¼º
À¯»ó¿ø
±èÇüÁÖ
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 22 NO. 02 PP. 0033 ~ 0047 (2006. 08) |
Çѱ۳»¿ë (Korean Abstract) |
ÃÖ±Ù ¿ÂÅç·ÎÁö°¡ Áö½Ä µµ¸ÞÀÎÀÇ Áö½ÄÀ» °øÀ¯Çϱâ À§ÇÑ ¼ö´ÜÀ¸·Î ¸¹ÀÌ »ç¿ëµÇ¸ç, ¿ÂÅç·ÎÁö ¿ë¾î¸¦ »ç¿ëÇÑ ÀÇ¹Ì Á¤º¸µµ ´Ã¾î³ª´Â Ãß¼¼ÀÌ´Ù. ÀÇ¹Ì Á¤º¸¶õ µ¥ÀÌÅÍ °´Ã¼°¡ °®´Â ¿©·¯ Àǹ̸¦ ¿ÂÅç·ÎÁö ¿ë¾î¸¦ »ç¿ëÇØ Ç¥ÇöÇÑ µ¥ÀÌÅ͸¦ ¸»ÇÑ´Ù. ÀÇ¹Ì Á¤º¸ÀÇ ÇÑ ¿¹·Î, ÃÖ±Ù ´ëºÎºÐÀÇ »ý¹°ÇÐ µ¥ÀÌÅͺ£À̽º¿¡¼ À¯ÀüÀÚ »ê¹°ÀÇ ±â´É¤·¸£ ³ªÅ¸³»±â À§ÇØ Gene Ontology(GO)ÀÇ ¿ë¾î¸¦ »ç¿ëÇÑ ÁÖ¼® Á¤º¸¸¦ µé ¼ö ÀÖ´Ù. ÀǹÌÀû À¯»ç¼º °Ë»öÀº ¿ÂÅç·ÎÁö¿Í ÀÇ¹Ì Á¤º¸¸¦ ÀÌ¿ëÇØ À¯»çÇÑ Àǹ̸¦ °®´Â °´Ã¼¸¦ °Ë»öÇÏ´Â ±â¹ýÀÌ´Ù. ÀüÅëÀûÀÎ À¯»ç¼º °Ë»ö°ú´Â ´Þ¸® ÀǹÌÀû À¯»ç¼º °Ë»öÀº ºñ ¸ÞÆ®¸¯ °ø°£¿¡ ´ëÇÑ °Ë»öÀÌ´Ù. º» ³í¹®¿¡¼´Â ÀÌ·¯ÇÑ ÀǹÌÀû À¯»ç¼º ó¸® ±â¹ý¿¡ ´ëÇØ Á¦¾ÈÇÑ´Ù. ¿ì¼±, ÀǹÌÀû À¯»ç¼º ÇÔ¼ö°¡ °®´Â ¿©·¯ ¼ºÁú¿¡ ´ëÇØ ¾Ë¾Æº¸°í À̸¦ ÀÌ¿ëÇÑ ÁúÀÇ Ã³¸®±â¹ýÀ» Á¦¾ÈÇÏ¿´´Ù. ºñ ¸ÞÆ®¸¯ °ø°£¿¡¼ÀÇ °Ë»öÀ» Ÿ¢ ¤À¿ª»öÀÎ ¸®½ºÆ®¸¦ »ç¿ëÇÑ ÁúÀÇ Ã³¸® ¹æ½ÄÀ» »ç¿ëÇϸç, °´Ã¼ÀÇ Å½»ö Ƚ¼ö¸¦ ÁÙÀ̱â À§ÇÑ ±â¹ýÀ» Á¦¾ÈÇÏ¿´´Ù. ½ÇÁ¦ÀÇ GO¿Í À¯ÀüÀÚ »ê¹°¿¡ ´ëÇÑ ÁÖ¼® Á¤º¸¸¦ »ç¿ëÇÑ ½ÇÇèÀ» ÅëÇØ Á¦¾ÈÇÑ ¾Ë°í¸®ÁòÀÌ È¿À²ÀûÀ̸ç È®À强À» °®°í ÀÖÀ½À» º¸¿´´Ù. |
¿µ¹®³»¿ë (English Abstract) |
As ontologies are widely used for sharing of domain knowledge, the amount of semantic information is also increasing. Semantic information of an object means the data which is represented with ontology terms to describe some properties of the object. The annotation data of Gene Ontology(GO) is a representative example of such semantic information, which describes the biological functionalities of gene products. Sematic similarity search is a technique that exploits the semantic information. Traditionally, similarity search considers the metric space. But the search space of semantic similarity search is non metric space. In this paper, we propose a query evaluation sheme for semantic similarity search. Firstly we investigate the properties of semantic similarity function and then we exploit these properties for semantic similarity search. For searching in non metric space, we use the inverted index lists and propose some methods to avoid useless accesses. Using GO and its annotation data as the real-life data, we show that the proposed method is efficient and scalabel. |
Å°¿öµå(Keyword) |
¿ÂÅç·ÎÁö
ÀǹÌÀû À¯»ç¼º °Ë»ö
Ontology
Sematic Similarity Search
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|