• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë

Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë

Current Result Document : 1 / 3   ´ÙÀ½°Ç ´ÙÀ½°Ç

ÇѱÛÁ¦¸ñ(Korean Title) ½º¸¶Æ®Æù »ç¿ëÀÚ À̵¿ °æ·Î¿¡ ´ëÇÑ È®·ü ±×·¡ÇÁ ¸ðµ¨ ÇнÀ ±â¹ý
¿µ¹®Á¦¸ñ(English Title) An Approach of Learning Probabilistic Graphical Models for Smartphone Users
ÀúÀÚ(Author) ±èÁ¦¹Î   ¹éÇýÁ¤   ¹Ú¿µÅà  Je-Min Kim   Hea-Jung Back   Young-Tack Park  
¿ø¹®¼ö·Ïó(Citation) VOL 41 NO. 02 PP. 0153 ~ 0163 (2014. 02)
Çѱ۳»¿ë
(Korean Abstract)
º» ³í¹®Àº ½º¸¶Æ®ÆùÀ¸·ÎºÎÅÍ ¼öÁýÇÑ gps Á¤º¸¸¦ ±â¹ÝÀ¸·Î »ç¿ëÀÚ °³ÀÎÀÇ °æ·Î ¸ðµ¨À» ÀÚµ¿À¸·Î ÇнÀÇÏ´Â ±â¹ý¿¡ ´ëÇØ ¼³¸íÇÑ´Ù. ½º¸¶Æ®ÆùÀÇ gps´Â ¼ö½ÅÀÌ ºÒ¾ÈÁ¤ÇÏ¿©, gpsÁÂÇ¥ ÀÌ·ÂÀ» ±â¹ÝÀ¸·Î »ç¿ëÀÚÀÇ ¸ðµç ¸ñÀûÁö³ª °æ·Î¸¦ Á¤È®ÇÏ°Ô ÆľÇÇϴµ¥ ¾î·Á¿òÀÌ µû¸¥´Ù. ÀÌ¿¡ º» ³í¹®Àº ¼¼ °¡Áö ºÎºÐ¿¡ ÃÊÁ¡À» ¸ÂÃß¾î ¼­¼úÇÏ¿´´Ù. ¸ÕÀú, »ç¿ëÀÚ °æ·Î Á¤º¸¸¦ È®·ü ¸ðµ¨·Î Ç¥ÇöÇÑ µ¿Àû º£ÀÌÁö¾È ¸Á¿¡ ´ëÇØ ¼³¸íÇÏ°í, µÎ ¹ø°, °ü½É ÁöÁ¡°ú À̵¿ ±¸°£ ±ºÁýÈ­ °úÁ¤À» ÅëÇØ Àüü ÈƷõ¥ÀÌÅ͸¦ Á¤»ó gps ±ËÀû°ú ºñÁ¤»ó gps ±ËÀûÀ¸·Î ±¸ºÐÇÑ ´ÙÀ½, Á¤»ó gps ±ËÀûÀ¸·ÎºÎÅÍ Ãʱ⠻ç¿ëÀÚ °æ·Î ¸ðµ¨À» ÃßÃâÇϱâ À§ÇÑ ¹æ¹ý¿¡ ´ëÇØ ¼Ò°³ÇÑ ÈÄ, ¸¶Áö¸·À¸·Î Ãʱ⠰æ·Î ¸ðµ¨À» ¹ÙÅÁÀ¸·Î ¿¹Ãø-ÃÖ´ëÈ­(expectation maximization) ¾Ë°í¸®ÁòÀ» Àû¿ëÇÏ¿© ÈƷõ¥ÀÌÅÍ Àüü¸¦ ´ë»óÀ¸·Î ½º¸¶Æ®Æù »ç¿ëÀÚÀÇ °³ÀÎ °æ·Î¿¡ ´ëÇÑ È®·ü ¸ðµ¨À» ÇнÀÇÏ´Â ¹æ¹ý¿¡ ´ëÇØ ¼³¸íÇÑ´Ù. º» ³í¹®¿¡¼­ Á¦¾ÈÇÏ´Â ¹æ½ÄÀ» ÅëÇØ ÇнÀµÈ °æ·Î ¸ðµ¨Àº »ç¿ëÀÚÀÇ ÇöÀç ¶Ç´Â ÇâÈÄ À§Ä¡³ª À̵¿ °æ·Î¸¦ È¿°úÀûÀ¸·Î ¿¹ÃøÇϱâ À§ÇÑ ÀÚ·á·Î Á¦°øµÇ±â ¶§¹®¿¡ ´Ù¾çÇÑ À§Ä¡ ±â¹Ý Áö´ÉÇü ¼­ºñ½º¿¡ ÀÌ¿ë °¡´ÉÇÏ´Ù
¿µ¹®³»¿ë
(English Abstract)
In this paper, we present a method for learning a personal route model from the GPS logs of a user's smartphone. The GPS signals of a smartphone are highly unstable; hence, it is difficult to correctly determine a user's destinations and routes from the GPS logs of his/her smartphone. We overcome this problem by adopting three approaches. First, a dynamic Bayesian network is used to represent user route information as a probabilistic model. Second, normal and abnormal GPS trajectories are distinguished from training data on the basis of points of interest (POIs) and trip clustering, and an initial route model is developed using the normal GPS trajectories. Finally, a probabilistic model for a smartphone user's personal routes is learned from training data by adapting the expectation-maximization algorithm (EM) based on the initial route model. The learned model can be applied to location-based intelligent services as it provides data for predicting a user's current (or future) locations and trips.
Å°¿öµå(Keyword) ½º¸¶Æ®Æù   À̵¿ °æ·Î ¸ðµ¨   µ¿Àû º£ÀÌÁö¾È ¸Á   ÆÄƼŬ ÇÊÅÍ   ¿¹Ãø-ÃÖ´ëÈ­ ¾Ë°í¸®Áò   smartphone   travel routes model   dynamic bayesian network   particle filter   dynamic bayesian network   particle filter  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå