• 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 ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) »ç¿ëÀÚÀÇ ÀÔ·Â Àǵµ¸¦ ¹Ý¿µÇÑ À½Àý N-gram ±â¹Ý Çѱ¹¾î ¶ç¾î¾²±â ¹× ºÙ¿©¾²±â ¿À·ù ±³Á¤ ½Ã½ºÅÛ
¿µ¹®Á¦¸ñ(English Title) A Korean Word Segmentation Error Correction System Reflecting User Intent based Syllable N-gram
ÀúÀÚ(Author) ¹Ú¼­¿¬   ¿Áö¿µ   Seoyeon Park   Cheolyoung Ock  
¿ø¹®¼ö·Ïó(Citation) VOL 27 NO. 03 PP. 0145 ~ 0150 (2021. 03)
Çѱ۳»¿ë
(Korean Abstract)
±âÁ¸ÀÇ ÀÚµ¿ ¶ç¾î¾²±â ½Ã½ºÅÛÀº »ç¿ëÀÚÀÇ ¶ç¾î¾²±â Á¤º¸¸¦ È°¿ëÇÏÁö ¾Ê°í ¶ç¾î¾²±â¸¦ ¸ðµÎ Á¦°ÅÇÑ ¹®Àå¿¡ ´ëÇØ °ø¹éÀ» »ðÀÔÇÏ´Â ¹æ½ÄÀ¸·Î ¶ç¾î¾²±â ¿À·ù¸¦ ¼öÁ¤ÇÑ´Ù. ÀÌ·¯ÇÑ ¹æ½ÄÀ¸·Î ¶ç¾î¾²±â ¿À·ù¸¦ ±³Á¤ÇÒ °æ¿ì, »ç¿ëÀÚ°¡ ¿Ã¹Ù¸£°Ô ÀÔ·ÂÇÑ ¶ç¾î¾²±â¸¦ ¼öÁ¤ÇÏ´Â ¹®Á¦¿Í »ç¿ëÀÚÀÇ Àǵµ¸¦ ÃæºÐÈ÷ ¹Ý¿µÇÏÁö ¸øÇÏ´Â ¹®Á¦°¡ ¹ß»ýÇÑ´Ù. º» ³í¹®¿¡¼­´Â ÀÌ·¯ÇÑ ¹®Á¦¸¦ º¸¿ÏÇϱâ À§ÇØ »ç¿ëÀÚ°¡ ÀÔ·ÂÇÑ Àǵµ¸¦ ¹Ý¿µÇÑ À½Àý N-gram ±â¹Ý Çѱ¹¾î ¶ç¾î¾²±â ¹× ºÙ¿©¾²±â ¿À·ù ±³Á¤ ½Ã½ºÅÛÀ» Á¦¾ÈÇÑ´Ù. ½ÇÇè °á°ú, ¿À·ù°¡ 10% Æ÷ÇÔµÈ ¹®Àå¿¡ ´ëÇؼ­ À½Àý ´ÜÀ§ Á¤È®·ü 99.05%, ¾îÀý ´ÜÀ§ F1 score 95.57%¶ó´Â ³ôÀº ¼º´ÉÀ» º¸¿´´Ù. ÀÌ´Â »ç¿ëÀÚÀÇ ¶ç¾î¾²±â Á¤º¸¸¦ È°¿ëÇÏÁö ¾ÊÀº ±âÁ¸ ¹æ½Äº¸´Ù À½Àý ´ÜÀ§ Á¤È®·ü 1.85%, ¾îÀý ´ÜÀ§ F1 score 5.84% Çâ»óµÈ °á°úÀÌ´Ù. ¶ÇÇÑ, µö·¯´× ¹æ½ÄÀÌ ¾Æ´Ñ À½Àý È®·ü Åë°èÁ¤º¸¸¸À» »ç¿ëÇÔÀ¸·Î½á ÃÊ´ç 2691.69 ¹®ÀåÀÇ ºü¸¥ ±³Á¤ ¼Óµµ¸¦ º¸¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Previous researches on automatic spacing corrected errors by inserting spaces in sentences without utilizing the user¡¯s spacing information. The present approach involves modifying the user¡¯s input incorrectly and a problem that does not sufficiently reflect user intent. In this paper, we propose a syllable N-gram based Korean word segmentation system that reflects the user¡¯s intent. The comparison between the proposed model and the model using previous methods demonstrated an increase in the syllable accuracy from 97.20% to 99.05% and the word F1 score from 89.73% to 95.57% in the proposed model. Also, the proposed model was able to correct 2691.69 sentences per second.
Å°¿öµå(Keyword) ÀÚµ¿ ¶ç¾î¾²±â   Çѱ¹¾î ¶ç¾î¾²±â ¹× ºÙ¿©¾²±â   »ç¿ëÀÚ Àǵµ ¹Ý¿µ ¶ç¾î¾²±â   À½Àý N-gram   automatic spacing   korean word segmentation   word segmentation reflecting user¡¯s intent   syllable n-gram  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå