• 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¸Ã³¸®ÇÐȸ ³í¹®Áö > Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ

Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) ½Ã°è¿­ ¿¹ÃøÀ» À§ÇÑ ½ºÅ¸ÀÏ ±â¹Ý Æ®·£½ºÆ÷¸Ó
¿µ¹®Á¦¸ñ(English Title) Style-Based Transformer for Time Series Forecasting
ÀúÀÚ(Author) ±èµ¿°Ç   ±è±¤¼ö   Dong-Keon Kim   Kwangsu Kim  
¿ø¹®¼ö·Ïó(Citation) VOL 10 NO. 12 PP. 0579 ~ 0586 (2021. 12)
Çѱ۳»¿ë
(Korean Abstract)
½Ã°è¿­ ¿¹ÃøÀº °ú°Å ½ÃÁ¡ÀÇ Á¤º¸¸¦ Åä´ë·Î ¹Ì·¡ ½ÃÁ¡ÀÇ Á¤º¸¸¦ ¿¹ÃøÇÏ´Â °ÍÀ» ¸»ÇÑ´Ù. ÇâÈÄ ½ÃÁ¡ÀÇ Á¤º¸¸¦ Á¤È®ÇÏ°Ô ¿¹ÃøÇÏ´Â °ÍÀº ´Ù¾çÇÑ ºÐ¾ß Àü·« ¼ö¸³, Á¤Ã¥ °áÁ¤ µîÀ» À§ÇØ È°¿ëµÇ±â ¶§¹®¿¡ ¸Å¿ì Áß¿äÇÏ´Ù. ÃÖ±Ù¿¡´Â Æ®·£½ºÆ÷¸Ó ¸ðµ¨ÀÌ ½Ã°è¿­ ¿¹Ãø ¸ðµ¨·Î¼­ ÁÖ·Î ¿¬±¸µÇ°í ÀÖ´Ù. ±×·¯³ª ±âÁ¸ÀÇ Æ®·£½ºÆ÷¸ÓÀÇ ¸ðµ¨Àº ¿¹Ãø ¼øÂ÷¸¦ Ãâ·ÂÇÒ ¶§ Ãâ·Â °á°ú¸¦ ´Ù½Ã ÀÔ·ÂÇÏ´Â ÀÚ°¡È¸±Í ±¸Á¶·Î µÇ¾î ÀÖ´Ù´Â ÇÑ°èÁ¡ÀÌ ÀÖ´Ù. ÀÌ ÇÑ°èÁ¡Àº ¸Ö¸® ¶³¾îÁø ½ÃÁ¡À» ¿¹ÃøÇÒ ¶§ Á¤È®µµ°¡ ¶³¾îÁø´Ù´Â ¹®Á¦Á¡À» ÃÊ·¡ÇÑ´Ù. º» ³í¹®¿¡¼­´Â ÀÌ·¯ÇÑ ¹®Á¦Á¡À» °³¼±ÇÏ°í ´õ Á¤È®ÇÑ ½Ã°è¿­ ¿¹ÃøÀ» À§ÇØ ½ºÅ¸ÀÏ º¯È¯ ±â¹ý¿¡ Âø¾ÈÇÑ ¼øÂ÷ µðÄÚµù ¸ðµ¨À» Á¦¾ÈÇÑ´Ù. Á¦¾ÈÇÏ´Â ¸ðµ¨Àº Æ®·£½ºÆ÷¸Ó-ÀÎÄÚ´õ¿¡¼­ °ú°Å Á¤º¸ÀÇ Æ¯¼ºÀ» ÃßÃâÇÏ°í, À̸¦ ½ºÅ¸ÀÏ-±â¹Ý µðÄÚ´õ¿¡ ¹Ý¿µÇÏ¿© ¿¹Ãø ½Ã°è¿­À» »ý¼ºÇÏ´Â ±¸Á¶·Î µÇ¾î ÀÖ´Ù. ÀÌ ±¸Á¶´Â ÀÚ°¡È¸±Í ¹æ½ÄÀÇ ±âÁ¸ÀÇ Æ®·£½ºÆ÷¸ÓÀÇ µðÄÚ´õ ±¸Á¶¿Í ´Ù¸£°Ô, ¿¹Ãø ¼øÂ÷¸¦ ÇѲ¨¹ø¿¡ Ãâ·ÂÇϱ⠶§¹®¿¡ ´õ ¸Õ ½ÃÁ¡ÀÇ Á¤º¸¸¦ Á» ´õ Á¤È®È÷ ¿¹ÃøÇÒ ¼ö ÀÖ´Ù´Â ÀåÁ¡ÀÌ ÀÖ´Ù. ¼­·Î ´Ù¸¥ µ¥ÀÌÅÍ Æ¯¼ºÀ» °¡Áö´Â ´Ù¾çÇÑ ½Ã°è¿­ µ¥ÀÌÅͼÂÀ¸·Î ¿¹Ãø ½ÇÇèÀ» ÁøÇàÇÑ °á°ú, º» ³í¹®¿¡¼­ Á¦½ÃÇÑ ¸ðµ¨ÀÌ ±âÁ¸ÀÇ ´Ù¸¥ ½Ã°è¿­ ¿¹Ãø ¸ðµ¨º¸´Ù ¿¹Ãø Á¤È®µµ°¡ ¿ì¼öÇÏ´Ù´Â °ÍÀ» º¸ÀδÙ.
¿µ¹®³»¿ë
(English Abstract)
Time series forecasting refers to predicting future time information based on past time information. Accurately predicting future information is crucial because it is used for establishing strategies or making policy decisions in various fields. Recently, a transformer model has been mainly studied for a time series prediction model. However, the existing transformer model has a limitation in that it has an auto-regressive structure in which the output result is input again when the prediction sequence is output. This limitation causes a problem in that accuracy is lowered when predicting a distant time point. This paper proposes a sequential decoding model focusing on the style transformation technique to handle these problems and make more precise time series forecasting. The proposed model has a structure in which the contents of past data are extracted from the transformer-encoder and reflected in the style-based decoder to generate the predictive sequence. Unlike the decoder structure of the conventional auto-regressive transformer, this structure has the advantage of being able to more accurately predict information from a distant view because the prediction sequence is output all at once. As a result of conducting a prediction experiment with various time series datasets with different data characteristics, it was shown that the model presented in this paper has better prediction accuracy than other existing time series prediction models.
Å°¿öµå(Keyword) ½Ã°è¿­ ¿¹Ãø   Æ®·£½ºÆ÷¸Ó   »ý¼º µðÄÚ´õ   ½ºÅ¸ÀÏ º¯È¯   Time Series Forecasting   Transformer   Generative Decoder   Style Transfer  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå