Á¤º¸°úÇÐȸ ³í¹®Áö D : µ¥ÀÌŸº£À̽º
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
¸ð¹ÙÀÏ ¼Ò¼È ³×Æ®¿öÅ©¸¦ À§ÇÑ »ç¿ëÀÚÀÇ ¼±È£µµ ¹× À̵¿ ÆÐÅÏÀ» ÀÌ¿ëÇÑ Ä£±¸ Ãßõ |
¿µ¹®Á¦¸ñ(English Title) |
Friend Recommendation Using User Preferences and Moving Patterns for Mobile Social Networks |
ÀúÀÚ(Author) |
ÀÌÃæÈñ
À¯Àç¼ö
º¹°æ¼ö
¹Ú¿ëÈÆ
ÀÓÁ¾ÅÂ
Chunghui Lee
Jaesoo Yoo
Kyoungsoo Bok
Yonghun Park
Jongtae Lim
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 40 NO. 01 PP. 0079 ~ 0087 (2013. 02) |
Çѱ۳»¿ë (Korean Abstract) |
ÃÖ±Ù »ç¿ëÀÚµéÀÇ ±ËÀû ºÐ¼®À» ÅëÇØ »ç¿ëÀÚÀÇ ¼ºÇâ¿¡ ÀûÇÕÇÑ Á¤º¸¸¦ ÃßõÇØ ÁÖ´Â ¿¬±¸µéÀÌ ÁøÇàµÇ°í ÀÖ´Ù. ÀÌ·¯ÇÑ ¿¬±¸µéÀº ¿©ÇàÁö Ãßõ, Ä£±¸ ã±â µî°ú °°Àº ÀÀ¿ë ¼ºñ½º¸¦ À§Çؼ Ŭ·¯½ºÅ͸µ ±â¹ý°ú ÆÐÅÏ ¸ÅĪ ±â¹ýÀ» ¸¹ÀÌ »ç¿ëÇÏ°í ÀÖ´Ù. ±×·¯³ª Ŭ·¯½ºÅ͸µ ±â¹ýÀº Ãßõ ¹Þ´Â »ç¿ëÀÚÀÇ ¼±È£µµ°¡ ¹Ý¿µµÇÁö ¾Ê°í, ´Ù¸¥ »ç¿ëÀÚµéÀÇ ¼±È£µµ¿¡ µû¶ó ÃßõÀ» ÇØÁÖ´Â ´ÜÁ¡ÀÌ Á¸ÀçÇÑ´Ù. ¶ÇÇÑ, ÆÐÅÏ ¸ÅĪ ±â¹ýÀº ´Ù¸¥ »ç¿ëÀÚ¿ÍÀÇ POI(Point of Interest)ÀÇ À¯Çü°ú °Å¸®¸¦ ºñ±³ÇÏ¿© ÃßõÀ» ¼öÇàÇϱ⠶§¹®¿¡ »ç¿ëÀÚÀÇ ¼¼ºÎÀûÀÎ ¼±È£µµ¸¦ ¹Ý¿µÇÒ ¼ö ¾ø´Â ´ÜÁ¡ÀÌ Á¸ÀçÇÑ´Ù. ÀÌ·¯ÇÑ ±âÁ¸ ¿¬±¸µéÀ» º¸¿ÏÇϱâ À§ÇØ º» ³í¹®¿¡¼´Â POIÀÇ ¼Ó¼º Á¤º¸¿Í »ç¿ëÀÚÀÇ À̵¿ ÆÐÅÏÀ» °í·ÁÇÑ Ä£±¸ Ãßõ ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. Á¦¾ÈÇÏ´Â ±â¹ýÀº Å©°Ô »ç¿ëÀÚÀÇ ¼Ó¼º Á¤º¸¸¦ ÀÌ¿ëÇؼ ¼±È£µµ¸¦ °è»êÇÏ¿© ¼±È£µµ°¡ ´Ù¸¥ »ç¿ëÀÚ¸¦ ÇÊÅ͸µÇÏ´Â ºÎºÐ°ú ÆÐÅÏ ¸ÅĪ ±â¹ýÀ» »ç¿ëÇÏ¿© »ç¿ëÀÚÀÇ ±ËÀû°ú ±ÙÁ¢ÇÏ°Ô À̵¿ÇÑ ±ËÀûÀ» ã´Â ºÎºÐÀ¸·Î ³ª´©¾îÁø´Ù. ¼º´É Æò°¡¸¦ ÅëÇؼ Ä£±¸·Î ÃßõµÈ »ç¿ëÀÚµéÀÇ POI ±ËÀû°ú »ç¿ëÀÚ POI ±ËÀûÀÌ À¯»çÇÔÀ» È®ÀÎÇÏ¿´°í, POI ±ËÀûÀÇ À¯»çµµ¸¦ ±âÁ¸ ±â¹ý°ú ºñ±³ÇÔÀ¸·Î½á Á¦¾ÈÇÏ´Â ±â¹ýÀÇ ¿ì¼ö¼ºÀ» ÀÔÁõÇÏ¿´´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Recently, it has been studied to recommend the proper information of user¡¯s tendency through the trajectory analysis of moving objects. The existing works often used clustering and pattern matching schemes for application services such as itinerary recommendation and finding friends. However, the clustering scheme has problems that it does not reflect the preference of users and recommend the information through the preference of other users. Since the pattern matching scheme is also performed by comparing the Point of Interests(POI) type and distance of other users, it does not reflect the detailed preference of users. We propose a friends recommendation system that considers the attributes of POIs and the moving patterns of users in order to alleviate the problems of the existing schemes. The proposed scheme is divided into two parts. One is to compute the preference using the attributes of users. Therefore, it filters users that have different preference. The other is to find the trajectories closer to the user¡¯s trajectory using the pattern matching scheme. In order to verify the superiority of the proposed scheme, it is clear through performance evaluation that it shows the high similarity of the recommended two moving patterns by the POI trajectories of users and user¡¯s trajectory. The proposed scheme outperforms the existing scheme in terms of the similarity of POI trajectories.
|
Å°¿öµå(Keyword) |
»ç¿ëÀÚ À¯»çµµ
POI ±ËÀû
À̵¿ ÆÐÅÏ
̵̧
User similarity
POI trajectory
Moving pattern
Recommendation
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|