µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Å°¿öµå ºÎÇÕ ½ºÄ«À̶óÀΠó¸® |
¿µ¹®Á¦¸ñ(English Title) |
On Processing Keyword-Matched Skylines |
ÀúÀÚ(Author) |
ÃÖÇö½Ä
¼º¹Î°æ
Á¤¿¬µ·
Hyunsik Choi
Min Kyoung Sung
Yon Dohn Chung
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 25 NO. 03 PP. 0113 ~ 0129 (2009. 12) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®¿¡¼´Â »õ·Î¿î ÇüÅÂÀÇ ½ºÄ«À̶óÀÎ ÁúÀÇÀÎ Å°¿öµå ºÎÇÕ ½ºÄ«À̶óÀÎ ÁúÀǸ¦ ¼Ò°³ÇÑ´Ù. Å°¿öµå ºÎÇÕ ½ºÄ«À̶óÀÎ ÁúÀÇ´Â ÅؽºÆ® Á¤º¸¸¦ Æ÷ÇÔÇÑ Â÷¿ø Æ©Çà ÁýÇÕ¿¡¼ ÁúÀÇ Å°¿öµå¸¦ °¡Áö´Â ÁýÇÕÀÇ ½ºÄ«À̶óÀÎÀ» ±¸ÇÏ´Â ¹®Á¦ÀÌ´Ù. ±âÁ¸ ½ºÄ«À̶óÀÎ ÁúÀÇó¸® ±â¹ýÀº ÅؽºÆ® °Ë»öÀ» ó¸®ÇÏÁö ¸øÇϱ⠶§¹®¿¡ ÀÌ¹Ì Ã£Àº Å°¿öµå ºÎÇÕ Æ©Çõé·ÎºÎÅÍ ´Ù½Ã ½ºÄ«À̶óÀÎÀ» ±¸ÇØ¾ß Çϸç ÀÌ´Â ¸¹Àº ºÎÇϸ¦ À¯¹ßÇÑ´Ù. µû¶ó¼ º» ³í¹®¿¡¼´Â Å°¿öµå ºÎÇÕ ½ºÄ«À̶óÀÎ ÁúÀǸ¦ À§ÇÑ È¿À²ÀûÀÎ ÁúÀÇó¸® ¾Ë°í¸®ÁòÀ» Á¦¾ÈÇÑ´Ù. Á¦¾È ¾Ë°í¸®ÁòÀº IR2-TreeÀ» »öÀÎ ±¸Á¶·Î »ç¿ëÇÏ°í ºÐ±âÇÑÁ¤¹ýÀ» ÀÌ¿ëÇØ Æ®¸®¸¦ ¼øȸÇÑ´Ù. µû¶ó¼ ¼øȸµ¿¾È Æ©ÇÃÀÇ º¤ÅÍ¿Í ÅؽºÆ® Á¤º¸¸¦ µÑ ´Ù °í·ÁÇÏ¿© ¹æ¹®ÇÒ ÇÊ¿ä°¡ ¾ø´Â ³ëµå´Â È¿À²ÀûÀ¸·Î °¡ÁöÄ¡±â ÇÑ´Ù. Á¦¾È ¾Ë°í¸®ÁòÀÇ ¼º´ÉÆò°¡¸¦ À§ÇØ ±âÀú ¾Ë°í¸®ÁòÀÎ ¿ª»öÀÎ ½ºÄ«À̶óÀÎ °Ë»ö±â¹ýÀ» Ãß°¡ÀûÀ¸·Î Á¦¾ÈÇÏ°í Å°¿öµå ºÎÇÕ ½ºÄ«À̶óÀÎ ÁúÀÇó¸® ±â¹ý°ú ºñ±³Æò°¡ ÇÑ´Ù. ¼º´ÉÆò°¡´Â ¿¬»ê ½Ã°£ ¹× I/Oºñ¿ë ¹× ÃÊÁ¡À» ¸ÂÃß¾î ¼öÇàµÇ¸ç ¼º´ÉÆò°¡ °á°ú´Â Á¦¾È ¾Ë°í¸®ÁòÀÌ ±âÀú ¾Ë°í¸®Áòº¸´Ù ¿ì¼öÇÏ´Ù´Â °ÍÀ» Àß º¸¿©ÁØ´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
In this paper, we introduce a new type of queries, keyword-matched skyline queries, that retrieve tuples which are not dominated by any other tuples and whose text description matches with the given query keywords. Since conventional skyline processing methods do not handle keyword matching, skyline query processing would be necessary on the keyword-matched tuples computed from n-tuples with textual information. However, this combination approach incurs a significant processing overhead. In this paper, we propose an efficient algorithm for keyword-matched skyline queries. Our proposed algorithm makes use of IR2-Tree as the index structure and branch and bound search to traverse the tree. Therefore, our proposed algorithm can prune unqualified nodes during traversing the tree. We carried out extensive experiments with various parameters and data sets. The results of the experiments show our proposed algorithm is very efficient with regard to computational cost and I/O cost.
|
Å°¿öµå(Keyword) |
½ºÄ«À̶óÀÎ
ÅؽºÆ® Á¤º¸ °Ë»ö
ÆÄ·¹Åä Æ©ÇÃ
µ¥ÀÌÅͺ£À̽º
Skyline
Text Information Retrieval
Pareto Tuple
Database
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|