• 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) ¼¼¹Ì-½ºÆ®¸² ¼­ºê½ÃÄö½º ¸ÅĪÀ» ÀÌ¿ëÇÑ ¿Â¶óÀÎ ¸Ó½Å·¯´×
¿µ¹®Á¦¸ñ(English Title) Online Machine Learning Using Semi-Stream Subsequence Matching
ÀúÀÚ(Author) ÀÌÁ¾ÇР  Jong-Hark Lee   ±èÈ«Áö   Hong-Ji Kim   À̱âÈÆ   Ki-Hoon Lee  
¿ø¹®¼ö·Ïó(Citation) VOL 36 NO. 02 PP. 0017 ~ 0027 (2020. 08)
Çѱ۳»¿ë
(Korean Abstract)
¿Â¶óÀÎ ¸Ó½Å·¯´×(online machine learning)Àº ½Ç½Ã°£ ½ºÆ®¸² µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÏ¿© Áö¼ÓÀûÀ¸·Î ¸ðµ¨À» °»½ÅÇÏ´Â ¸Ó½Å·¯´× ¹æ¹ýÀÌ´Ù. ÃÖ±Ù ½Ã°è¿­ µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÑ ¿Â¶óÀÎ ¸Ó½Å·¯´× ¿¬±¸°¡ ²ÙÁØÈ÷ Áõ°¡ÇÏ°í ÀÖÁö¸¸, ¼¼¹Ì-½ºÆ®¸²(semi-stream) ¿¬»êÀ» Àû¿ëÇÑ ¿¬±¸´Â ¾ÆÁ÷ ºÎÁ·ÇÏ´Ù. ¼¼¹Ì-½ºÆ®¸² ¿¬»êÀº ½ºÆ®¸² µ¥ÀÌÅÍ¿¡ µð½ºÅ©¿¡ ÀúÀåµÈ µ¥ÀÌÅ͸¦ °áÇÕÇÏ´Â ¿¬»êÀ¸·Î, ½ºÆ®¸² µ¥ÀÌÅÍ¿Í À¯»çÇÑ ÆÐÅÏÀ» °¡Áö´Â µ¥ÀÌÅ͸¦ ÇÔ²² °áÇÕÇϱâ À§ÇØ ÇÊ¿äÇÑ ¿¬»êÀÌ´Ù. º» ³í¹®¿¡¼­´Â ½Ã°è¿­ µ¥ÀÌÅÍ¿¡ ´ëÇÑ ¿Â¶óÀÎ ¸Ó½Å·¯´×ÀÇ ¿¹Ãø ¼º´ÉÀ» ³ôÀ̱â À§ÇØ ¼¼¹Ì-½ºÆ®¸² ¿¬»êÀ» Àû¿ëÇÏ°í, µÎ µ¥ÀÌÅ͸¦ °áÇÕÇÏ´Â Á¶°ÇÀ¸·Î ¼­ºê½ÃÄö½º ¸ÅĪ(subsequence matching) ¿¬»êÀ» ÀÌ¿ëÇÑ´Ù. ½ÇÁ¦ ½Ã°è¿­ µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÏ¿© ½ÇÇèÇÑ °á°ú, ½ºÆ®¸² µ¥ÀÌÅ͸¸À» ÀÌ¿ëÇÏ´Â ¹æ¹ý¿¡ ºñÇØ Á¦¾ÈÇÑ ¹æ¹ýÀÇ Æò±ÕÁ¦°ö¿ÀÂ÷(MSE)°¡ Æò±Õ 6.61% °¨¼ÒÇÔÀ» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Online machine learning is a machine learning method that continuously updates models using the real-time stream data. Although online machine learning research using the time-series data has been steadily increasing, little attention has been paid to a semi-stream operation. A semi-stream operation combines the real-time data with disk-based stored data, which is required to combine the stored data having a pattern similar to the stream data. In this paper, a semi-stream operation is applied to enhance the prediction performance of the online machine learning, and a subsequence matching operation is used to combine the stream data with the stored data. The experimental results using a real-world time series dataset show that the proposed method reduces the average mean square error (MSE) by 6.61% compared with the method using only the stream data.
Å°¿öµå(Keyword) ¿Â¶óÀÎ ¸Ó½Å·¯´×   ¼­ºê½ÃÄö½º ¸ÅĪ   ¼¼¹Ì-½ºÆ®¸²   ½Ã°è¿­ µ¥ÀÌÅÍ   online machine learning   subsequence matching   semi-stream   time-series data  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå