µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
´ÙÁß ½ºÆ®¸²¿¡¼ È¿À²ÀûÀÎ »ó°ü°ü°è ±×·¡ÇÁ ¸¶ÀÌ´× |
¿µ¹®Á¦¸ñ(English Title) |
Efficient Correlated Graph Mining from Multiple Streams |
ÀúÀÚ(Author) |
±èÇö¿í
¹Ú±â¼º
Çѿ뱸
ÀÌ¿µ±¸
Hyunwook Kim
Kisung Park
Yongkoo Han
Young-Koo Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 31 NO. 03 PP. 0003 ~ 0013 (2015. 12) |
Çѱ۳»¿ë (Korean Abstract) |
±×·¡ÇÁ ½ºÆ®¸²¿¡¼ »ó°ü°ü°è ÆÐÅÏÀº ´Ù¾çÇÑ ºÐ¾ß¿¡¼ À¯¿ëÇÑ Áö½ÄÀ¸·Î È°¿ëµÉ ¼ö ÀÖ´Ù. ÃÖ±Ù, ´ÜÀÏ ±×·¡ÇÁ ½ºÆ®¸² ȯ°æ¿¡¼ »ó°ü°ü°è ÆÐÅÏÀ» È¿À²ÀûÀ¸·Î ã±â À§ÇÑ ´Ù¾çÇÑ ±â¹ýµéÀÌ Á¦¾ÈµÇ¾ú´Ù. ±×·¯³ª ±âÁ¸ÀÇ ±â¹ýµéÀº ´ÙÁß ±×·¡ÇÁ ½ºÆ®¸² ȯ°æ¿¡¼´Â ¸Å¿ì ¸¹Àº ¼öÀÇ ºÎºÐ±×·¡ÇÁ µ¿Çü °Ë»ç¿Í °°Àº ºñÈ¿À²ÀûÀÎ ÇÁ·Î¼¼½ºµéÀ» °¢ ½ºÆ®¸²¸¶´Ù ¹Ýº¹ÀûÀ¸·Î ¼öÇàÇØ¾ß ÇϹǷΠ¿À·£ ¼öÇà ½Ã°£À» ¿ä±¸ÇÑ´Ù. º» ³í¹®¿¡¼´Â ´ÙÁß ½ºÆ®¸² ȯ°æ¿¡¼ È¿À²ÀûÀÎ »ó°ü°ü°è ±×·¡ÇÁ ¸¶ÀÌ´× ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. Á¦¾ÈÇÏ´Â ±â¹ýÀº °¢ ½ºÆ®¸²¿¡¼ ¼öÇàÇÑ ºó¹ß ºÎºÐ±×·¡ÇÁ ¸¶ÀÌ´×ÀÇ Å½»öÆ®¸®¸¦ ÀÌ¿ëÇÏ¿© È¿À²ÀûÀ¸·Î ºÎºÐ±×·¡ÇÁ µ¿Çü °Ë»ç¸¦ ¼öÇàÇÑ´Ù. ¶ÇÇÑ, °ø°£ »ç¿ë·®À» °³¼±Çϱâ À§ÇÑ Å½»öÆ®¸® º´ÇÕ ±â¹ýµµ Á¦¾ÈÇÑ´Ù. ½ÇÇèÀ» ÅëÇØ Á¦¾ÈÇÏ´Â ±â¹ýÀÌ ±âÁ¸ÀÇ ´ÜÀÏ ½ºÆ®¸² ±â¹Ý ±â¹ýº¸´Ù ¼öÇà½Ã°£ÀÌ ¾à 10~20%Çâ»óµÊÀ» º¸ÀδÙ.
|
¿µ¹®³»¿ë (English Abstract) |
Correlated patterns from graph streams can be utilized informative knowledges in various applications. Recently, various approaches for mining correlated graphs from single graph streams are proposed. However, these approaches require long running time in multiple stream environments because inefficient processes such as a large number of subgraph isomorphism tests must be iteratively performed for each graph stream. In this paper, we propose an efficient correlated graph mining approach from multiple streams. The proposed approach perform subgraph isomorphism by using the searching tree of frequent pattern mining which is performed by correlated pattern mining. Moreover, we also propose the tree merging technique for optimizing space usage. In experiment, we show that the proposed approach can reduce execution time by up to 10~20% compared with the existing single stream based correlated graph mining method.
|
Å°¿öµå(Keyword) |
»ó°ü°ü°è ±×·¡ÇÁ
´ÙÁß ½ºÆ®¸² ó¸®
±×·¡ÇÁ ½ºÆ®¸²
Correlated graph
Multiple stream processing
Graph stream
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|