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

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ÇÐȸÁö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ÇÐȸÁö > µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) ½º¸¶Æ®Æù »ç¿ëÀÚÀÇ Áß´Ü°¡´É¼º ¿¹Ãø ½Ã ÄÝµå ½ºÅ¸Æ® ¹®Á¦ ÇØ°á ¿¬±¸
¿µ¹®Á¦¸ñ(English Title) On Mitigating the Cold Start Problem in Predicting Interruptibility of Smartphone Users
ÀúÀÚ(Author) ¾çÇö¼®   Ãֹμö   ÀÌÀç±æ   HyunSeok Yang   Minsoo Choy   Jae-Gil Lee  
¿ø¹®¼ö·Ïó(Citation) VOL 32 NO. 03 PP. 0023 ~ 0034 (2016. 12)
Çѱ۳»¿ë
(Korean Abstract)
½º¸¶Æ®Æù ½ÃÀåÀÇ ¼ºÀåÀº ¿¬±¸ÀÚµé·Î ÇÏ¿©±Ý »ç¿ëÀÚµéÀÇ »óȲ (Context) Æľǰú ±×¿¡ ±â¹Ý ÇÑ Ãßõ¿¡ °ü·ÃµÈ ¿¬±¸¸¦ ¼ö¿ùÇÏ°Ô ÇØÁÖ¾ú´Ù. ÇÏÁö¸¸ »ç¿ëÀÚÀÇ »óȲÀ» ÀÎÁöÇÏ°í Áß´Ü°¡´É¼ºÀ» ÆÇ´ÜÇÏ´Â °ÍÀº ½¬¿î ÀÏÀÌ ¾Æ´Ï´Ù. ƯÈ÷ ½Å±Ô »ç¿ëÀÚ¿Í °°ÀÌ Ãʱ⠵¥ÀÌÅÍ°¡ ºÎÁ·ÇÑ »óȲ¿¡´Â ´õ¿í ¾î·Æ´Ù. ÀÌ¿¡, ¾Õ¼± ¿¬±¸¿¡¼­´Â ½Å±Ô »ç¿ëÀÚ¿Í À¯»çÇÑ ÇൿÆÐÅÏÀ» º¸ÀÌ´Â ±âÁ¸ »ç¿ëÀÚµéÀ» Ãß°¡ÇÏ¿© ¿¹Ãø Á¤È®µµ¸¦ ³ôÀÌ´Â ¹æ¹ýÀ» Á¦¾ÈÇÏ¿´Áö¸¸ ÇÑ°è°¡ ÀÖ¾ú´Ù. º» ¿¬±¸¿¡¼­´Â ½Å±Ô »ç¿ëÀÚÀÇ Áß´Ü°¡´É¼º ¿¹Ãø¸ðµ¨ÀÇ Á¤È®µµ¸¦ ´õ ³ôÀ̱â À§ÇØ ¼¼ °¡Áö ´Ù¸¥ ¹æ¹ýÀ» ½Ãµµ ÇÏ¿´´Ù. ù°·Î, À¯»ç »ç¿ëÀÚµéÀ» ÆǺ°ÇÏ´Â °úÁ¤¿¡¼­ »ç¿ëµÇ´Â °¡ÁßÄ¡ °ªµéÀ» Á¶Á¤ÇÏ¿´´Ù. µÑ°·Î´Â ±âÁ¸ »ç¿ëÀÚ¿Í À¯»çµµ¸¦ °è»êÇÏ°í Á¤ÇØÁø ±âÁØ¿¡ µû¶ó »ç¿ëÀÚ¸¦ Ãß°¡ÇÏ´Â ¹æ½ÄÀ» Á¦¾ÈÇÏ¿´´Ù. ¼Â°·Î´Â ±âÁ¸ »ç¿ëÀÚ¿Í À¯»ç »ç¿ëÀÚÀÇ µ¥ÀÌÅ͸¦ ÅëÇÕÇÏ¿© ÇϳªÀÇ ¿¹Ãø ¸ðµ¨À» »ý¼ºÇÏ´Â ¹æ½Ä ´ë½Å¿¡ °¢ »ç¿ëÀÚÀÇ µ¥ÀÌÅ͸¦ È°¿ëÇÏ¿© °¢°¢ÀÇ ¿¹Ãø ¸ðµ¨À» ¸¸µé°í °á°ú¸¦ ¿¹ÃøÇÏ¿© Á¾ÇÕÇÏ´Â ¹æ½ÄÀ» Àû¿ëÇÏ¿´´Ù. ´Ù¾çÇÑ ¿¹Ãø¸ðµ¨ »ý¼º¹æ¹ýÀ» ÅëÇØ Ãʱ⠵¥ÀÌÅÍ È®º¸°¡ ¾î·Á¿î ½º¸¶Æ®Æù ½Å±Ô »ç¿ëÀÚÀÇ Áß´Ü°¡´É¼º ¿¹Ãø Á¤È®µµ¸¦ Çâ»ó½ÃÅ°´Â ¹æ¹ýÀ» Á¦½ÃÇÑ´Ù.
¿µ¹®³»¿ë
(English Abstract)
The Growth of smartphones has made researchers to do a study related to context-aware recommendation area easily. However, predicting interruptibility by recognizing user¡¯s context is not easy. Especially, prediction for a new user is more difficult due to the lack of available information. In previous study, we suggested a method of identifying similar users with respect to the user behavior of the new user and using their information for the interruptibility prediction. We achieved higher accuracy but it was not that significant. In this paper, we tried three ways to obtain better performance. First, we manipulated weights which are used for the similar user computation precess. Second, we changed the number of similar users to utilize by setting a similarity score threshold. Third, instead of model aggregation, we consider individual prediction result of models of the new user and similar users, and make a final decision. Like this, we propose various ways to tackle the cold start problem in interruptibility prediction.
Å°¿öµå(Keyword) Áß´Ü°¡´É¼º   »óȲÀÎÁö   ½º¸¶Æ®Æù   ÄÝµå ½ºÅ¸Æ® ¹®Á¦   Interruptibility   Context-Awareness   Smartphone   Cold Start Problem  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå