Á¤º¸°úÇÐȸ ³í¹®Áö D : µ¥ÀÌŸº£À̽º
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
±ËÀû ±×·¡ÇÁ ÁýÇÕ À¯»çµµ ÃøÁ¤ ±â¹ý |
¿µ¹®Á¦¸ñ(English Title) |
A Method for Measuring Similarity between Trajectory Graph Sets |
ÀúÀÚ(Author) |
È«ÁöÇý
¹Ú±â¼º
Çѿ뱸
ÀÌ¿µ±¸
Jihye Hong
Kisung Park
Yongkoo Han
Young-Koo Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 40 NO. 03 PP. 0153 ~ 0158 (2013. 06) |
Çѱ۳»¿ë (Korean Abstract) |
GPS ¼¾¼°¡ ³»ÀåµÈ ½º¸¶Æ®ÆùÀÇ ´ëÁßÈ¿¡ µû¶ó ±ËÀû µ¥ÀÌÅÍ µ¥ÀÌŸº£À̽º ±¸ÃàÀÌ ¿ëÀÌÇØÁ³´Ù. ÃÖ±Ù °³ÀÎÀÇ »ýÈ°ÆÐÅÏÀ» ¹Ý¿µÇÒ ¼ö ÀÖ´Â ±×·¡ÇÁ ±â¹Ý ±ËÀû µ¥ÀÌÅÍ ¸ðµ¨¸µ ¹æ¹ýÀÌ Á¦¾ÈµÇ¾ú´Ù. ±×·¯³ª ÀÌ ¿¬±¸´Â ±ËÀû µ¥ÀÌÅÍ ¸ðµ¨¸µ ¹æ¹ýÀ» ÁÖ·Î Á¦¾ÈÇÏ¿©, °³ÀÎÈ ¼ºñ½º¿Í °°Àº ÀÀ¿ë ºÐ¾ß¿¡ »ç¿ëÇÒ ¼ö ÀÖ´Â ¸¶ÀÌ´× ±â¹ýµéÀº Á¦¾ÈÇÏÁö ¾Ê¾Ò´Ù. º» ³í¹®¿¡¼´Â ±ËÀû ±×·¡ÇÁ ÁýÇÕÀ¸·Î Ç¥ÇöµÇ´Â »ç¿ëÀÚµé °£ÀÇ À¯»çµµ¸¦ È¿°úÀûÀ¸·Î ÃøÁ¤ÇÏ´Â ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. Á¦¾ÈÇÏ´Â À¯»çµµ ÃøÁ¤ ±â¹ýÀº »ç¿ëÀÚ¸¶´ÙÀÇ °íÀ¯ÇÑ »ýÈ°ÆÐÅÏ Æ¯Â¡À» Àß ¹Ý¿µÇÒ ¼ö ÀÖ´Â ´ëÇ¥ ºó¹ß ºÎºÐ±×·¡ÇÁµéÀ» ã¾Æ À¯»çµµ¸¦ ºñ±³ÇÑ´Ù. À¯»çµµ¸¦ È¿°úÀûÀ¸·Î °è»êÇϱâ À§ÇÏ¿©, ÁýÇÕ °£ÀÇ °Å¸® ÃøÁ¤ ¾Ë°í¸®ÁòÀÎ Hausdorff °Å¸®¿Í µÎ ±×·¡ÇÁ°£ÀÇ À¯»çµµ ÃøÁ¤ ¾Ë°í¸®ÁòÀÎ ÃÖ´ë °øÅë ºÎºÐ±×·¡ÇÁ¸¦ ÀÌ¿ëÇÑ ±×·¡ÇÁ ÁýÇÕ À¯»çµµ ÃøÁ¤ ¾Ë°í¸®ÁòÀ» Á¦¾ÈÇÑ´Ù. ½ÇÇèÀ» ÅëÇÏ¿© Á¦¾ÈÇÏ´Â À¯»çµµ ÃøÁ¤ ±â¹ýÀÌ »ç¿ëÀÚ °£ÀÇ À¯»çµµ¸¦ È¿°úÀûÀ¸·Î ÃøÁ¤ÇÒ ¼ö ÀÖÀ½À» º¸ÀδÙ. |
¿µ¹®³»¿ë (English Abstract) |
As a number of people use smart phones with embedded GPS sensors, it becomes easy to construct a trajectory data database. Recently, a graph based trajectory modeling study has been performed, which can reflect personal lifestyles. However, the study mainly has focused on a modeling method but not suggested mining techniques that can be used for applications such as personalized services. In this paper, we propose a method that measures a similarity between users represented by sets of trajectory graphs. The proposed method measures the similarity between users' feature frequent subgraphs, which imply each user's essential lifestyles, rather than trajectory graphs. In order to effectively calculate the similarity, we propose a graph set similarity algorithm using Hausdorff distance for calculating a set similarity and maximum common subgraph for calculating a graph similarity. In the experiment, we show our proposed method can measure similarities between users effectively. |
Å°¿öµå(Keyword) |
±ËÀû µ¥ÀÌÅÍ
±×·¡ÇÁ ¸¶ÀÌ´×
ÁýÇÕ À¯»çµµ
trajectory data
graph mining
set similarity
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|